SCIENCE & RESEARCH Traffic planning International Transportation | Collection 2021 50 Assessment of autonomous moving vehicles From theoretical approaches to practical test-procedures Car-inherent capabilities, Conditioning driving style, Traffic presence, Complexity handling, Testing arrangements Irrespective of deployment strategies of the automotive sector a consistent procedure of testing and proving automatization technologies is required. System adaptions in technical respect and coexistence strategies with view on traffic practice will be necessary. The multitude of motorized road users and their physical capabilities to pass interactions frictionless is challenging. That consideration leads to questions of two kinds, firstly what knowledge is needed for developing the automat system and secondly how transparent the algorithmic conditioning will be handled by the car suppliers. Heinz Doerr, Viktoria Marsch, Andreas Romstorfer F irst of all, irrespective of market deployment strategies of the automotive sector a chained up procedure of testing and proving automatization technologies is required. One can assume that these innovations would not change our mobility needs and the existent transport systems completely. But then mutual system adaptions in technical respect and coexistence strategies with view on traffic practice will be necessary for an utmost frictionless implementation of different automated vehicles onto the road-network. Research in view of the multitude of motorized road users and their technical capabilities to pass interactions without any conflict with others is challenging. It claims a decisive control mastering hidden in the backend of the car-inherent automatic chain. That consideration leads to questions of two kinds, firstly what knowledge is required for developing, testing and proving as well as secondly how transparent the algorithmic conditioning will be handled by the car suppliers. It concerns car-holders and other groups of road users as well as bearers of public road infrastructure. Knowledge about traffic events One can dare the assumption that there would be a lack of knowledge about interrelations between traffic-participants with the view on automated traffic operations. Systematic observations of the traffic practised onto our road network nowadays could deepen such an understanding. To this end to put up a framework of systematic orders describing the manifold mobility groups might help. Each of them - motorized or not - are characterized in respect of their specific range of traffic behavior due to its movement capabilities. In such a way logical procedures for developing and testing of automatic functionalities could be established 1 . It facilitates probably licensing by authorities and makes implementation onto road network more transparent for political decision makers. Finally the degree of acceptance in public could be fostered if a spatial zoning of regulations will have to be ordered by municipalities like in pedestrian’s or residential zones of our habitat. Such an approach represents necessarily a questioning view on that technological progress complementary to product promotion. It concerns besides deployment on car-buyer’s markets by the automobile manufactures all questions of implementation into the mobility system. But with it methodical approaches of gaining recognition about traffic events and eventualities should not only focused on interrelations between motorists driving on their playing-ground roadway 2 . Although this field of knowledge is rarely cultivated so far except for traffic accidence research. Expectations with regard to utility of automatizing traffic let recognize research tasks to deal with and technical development problems to solve. At very first, the behavior of traffic participants observed as physical moving bodies onto road spaces serves as resource of knowledge aiming at an implementation of automated road traffic in a consensual way. Such observations generate pictures of incidental scenes which can be communicated well. Furthermore a permanent traffic monitoring delivers data about frequencies and variations of road users interacting on the scenery. A systematic categorization of phenotypical sceneries, interaction-scenes and involved actor-groups depicted within a framework of orders and matrices makes complex interrelations crosswise visible and traceable. In that way scenarios for testing arrangements and for settings to realize them can be prepared. 3 Such fundamental knowledge should inspire a discourse about goals with respect to traffic safety, data Traffic planning SCIENCE & RESEARCH International Transportation | Collection 2021 51 security, desirable use cases, favourable utility areas - one should not ignore the spatial dimension of innovation’s diffusion - and environmental effects concerning energy reduction and emission relief for more air quality. A question connected is who bears the additional costs for equipment installed in cars and for investments resp. maintenance along the roadways, like for 5G/ 6G communication infrastructure. Finally current costs due to service charges for the road users as customers would have to be enquired too. Foremost, if one surveys relevant literature an extended demand of knowledge about trivial but complex traffic phenomena seems to be needed for anticipating future scenarios of mobility using road-net. Such an inquiring and reflecting discourse may deal with topics as follows: •• Theoretical fundamentals to understand relevant phenomena of daily traffic operations •• Methodical approaches finding crucial points of development in respect of technology assessment •• Multidisciplinary view on consequences of automatization of vehicles in respect of utility •• Testing procedures to prove technological functionality and to estimate affections on other road users •• Implementation strategies into mobility system and road network as political tasks •• Monitoring of practice onto road and in public spaces Car-inherent capabilities and external traffic conditions Theoretical approaches and methodical tooling to prepare testing and proving arrangements have two dimensions of viewing on challenges caused by traffic phenomena and technological effects: 1. The internal view on car-inherent capabilities as preconditions for automatization of driving (under the remaining responsibility of a driver) and for autonomization of car’s moving on public road spaces in future (at which responsibilities have to be clarified still). It is the preferred approach by automotive research and development, at which the motorization of a car type and the driver’s presence in a single vehicle build up the performing system. To that the perception of surroundings to detect opponents and obstacles to find open trajectory spaces in order to realize the best of trajectory options without risking conflicts will be essential. 2. The external view on moving cars (like a bird’s or drone’s view) as members of here so-called coincidental collectives sailing onto the roadway 4 . It is the neglected vision hardly anybody will know neither automotive developers nor traffic planners. That concerns interrelations between different motorized vehicles as well as with other road users. Moreover the view is widened on exogenous frame conditions by the road network in respect of topography and capacity. Additional we have to consider opportunities which are given by using interconnectivity for traffic actions. That may be realized in an interactive manner as data exchange between vehicles to vehicles (vs2vs) or furthermore as an information flow vehicles to infrastructure in all its service and capacity functions (vs2if ). To round off complexity time-depending traffic conditions have to be taken into account (see figure 2). In such a way fields of scenes for one and the same roadway section (= edge) can be composed before scenarios to prepare test arrangements could be constructed (see figures 7-9). Vehicle’s potential of motor power and conditioning vehicle’s driving style as moving pattern Each vehicle seen as a moving body manifests its individual resp. mark-typical behavior onto the roadway. That depends car-inherent on the available driving potentials for a classified car which is automated on different levels as depicted in figure 1 and external on the actual traffic conditions as drafted in figure 2. In the long term view driving modes would be distinguished in twice respects: As driving style resp. type of traffic behavior on the one hand and as traffic presence of a vehicle moving on a roadway on the other hand. The former depends either on the individual driving style of a human driver (as usual nowadays) or on the conditioned moving style of the automatic chain of a vehicle type equipped on a higher SAE-Level 5 . As second the traffic presence of a vehicle describes the operative mode either in dependence on individual decision making of a solitary moving car or on an external determined resp. associated moving pattern in considerateness with H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 3 achieved if a remote controlling by an exogenous control mastering elsewhere will take over. At the end of the day that would mean cars will be operating at an ultimate stage of “deautonomiziation”. Figure 1: Vehicle´s potential for autonomization connecting power performance with automation functions Car-inherent power performance Traffic-relevant autonomization potential Car-specific powertrain-equipment Capability for scenario-prediction to claim trajectory Vehicle-type as offered to customers Grade of exploitation of potential for driving dynamics Intensity of interconnectivity whilst interacting with other cars on the roadway Draft by H. Doerr (2020) All figures and images by the authors Driver´s presence as human-machine interaction The solitary traffic behavior of a human driven or automated moved vehicle succumbs on individual motivations which can be various according to day-time and trip purposes (like commuting or services needed just-in-time). But the moving potential of a vehicle due to its propulsion powerfulness has to be taken into account. Besides that, automation functionalities on different system levels installed in the car are more and more intervening and maybe overruling the individual human driver´s steering. In the end algorithms will take over decisively car´s movements tracking on the roadway. There, it would be a match between motorized vehicles distinguished by its “powerfulness” and by SAE-levels of automatization. In cases of autonomous moving vehicles (generally driverless or during driver´s absence of mind on board) it would be a crucial point which car of the coincidental mixture of vehicles would succeed and which of them would yield (= give way) when interactions take place. It traces back to the question how a car-inherent autonomization-system had been conditioned by the automobile manufacturer. But that is not enough at all if facing daily traffic conditions. Vehicle´s traffic presence as precondition An external air-view on moving vehicles over an edge of the roadway provokes the question in which manner a vehicle is directed on his pathway (compare Fig. 8). The control master could be located either car-internal by the automatic chain on board or car-external by a control mastering of its movements located at a proprietary fleet-disposition center resp. at a central road traffic management. That means in a long-term perspective to clear up the future modes of traffic presence of each single vehicle onto the road-network which can be distinguished as follow:  Solitary traffic presence as an individual (inter)acting moving body. That could be operated as steering by a driver as usual but it could be also a robot-vehicle seeking his trajectory automatically.  Collective traffic presence as vehicle interconnected with adjacent cars which behaves with solidarity. Interactions would be evoked either by a defensive driving style of a driver or by a programed conditioning of the car-inherent automated controlling chain.  Traffic presence by remote steering a vehicle. That could be operated by an infrastructure-provider as capacity-oriented traffic flow optimization in designated traffic regions like in inner-towns or in networks of expressways. A remote controlling of single moving cars or a fleet of service vehicles, like robot-cabs, frequenting all over the road-network seems not so promising if one faces the mixture of different equipped vehicles and other road using groups. Automation-Level (SAE) as vehicle´s specific equipment with such typical functions efficient (a1..-n) Usage according to vehicle-classification and driving license (like P = Passenger car) P 2a4 (c2, t2) Conditioning of driving-style as program to shift (c1-n or c1+2+n) Mode of traffic presence according to stages of dependence (t1-n) Powerfulness in respect of gross mass of the vehicle and driving-relevant parameters (P 1..-n ) Figure 1: Vehicle’s potential for autonomization connecting power performance with automation functions SCIENCE & RESEARCH Traffic planning International Transportation | Collection 2021 52 other adjacent cars. But the last level would be achieved if a remote controlling by an exogenous control mastering elsewhere will take over. At the end of the day that would mean cars will be operating at an ultimate stage of “deautonomiziation”. Driver’s presence as human-machine interaction The solitary traffic behavior of a human driven or automated moved vehicle succumbs on individual motivations which can be various according to day-time and trip purposes (like commuting or services needed justin-time). But the moving potential of a vehicle due to its propulsion powerfulness has to be taken into account. Besides that, automation functionalities on different system levels installed in the car are more and more intervening and maybe overruling the individual human driver’s steering. In the end algorithms will take over decisively car’s movements tracking on the roadway. There, it would be a match between motorized vehicles distinguished by its “powerfulness” and by SAE-levels of automatization. In cases of autonomous moving vehicles (generally driverless or during driver’s absence of mind on board) it would be a crucial point which car of the coincidental mixture of vehicles would succeed and which of them would yield (= give way) when interactions take place. It traces back to the question how a carinherent autonomization-system had been conditioned by the automobile manufacturer. But that is not enough at all if facing daily traffic conditions. Vehicle’s traffic presence as precondition An external air-view on moving vehicles over an edge of the roadway provokes the question in which manner a vehicle is directed on his pathway (compare figure 8). The control master could be located either car-internal by the automatic chain on board or car-external by a control mastering of its movements located at a proprietary fleet-disposition center resp. at a central road traffic management. That means in a long-term perspective to clear up the future modes of traffic presence of each single vehicle onto the road-network which can be distinguished as follow: • Solitary traffic presence as an individual (inter)acting moving body. That could be operated as steering by a driver as usual but it could be also a robot-vehicle seeking his trajectory automatically. • Collective traffic presence as vehicle interconnected with adjacent cars which behaves with solidarity. Interactions would be evoked either by a defensive driving style of a driver or by a programed conditioning of the car-inherent automated controlling chain. • Traffic presence by remote steering a vehicle. That could be operated by an infrastructure-provider as capacity-oriented traffic flow optimization in designated traffic regions like in inner-towns or in networks of expressways. A remote controlling of single moving cars or a fleet of service vehicles, like robot-cabs, frequenting all over the road-network seems not so promising if one faces the mixture of different equipped vehicles and other road using groups. The hypothetical drafted diagram (figure 2) visualizes traffic relevant factors which constitute the changeable frame conditions for interactions on the roadway. The key mechanism of traffic flow is represented through the indicator of “Level of Service” (in stages A till E) describing the quality of traffic conditions in respect of driving options for each vehicle onto a roadway edge 6 . That depends on the density of vehicles present on an observed section and is determined through the mixture of vehicle types. Such circumstances enable or make it more difficult for vehicles to seek for open trajectory spaces. Both influence the velocity of the traffic flow. Moreover it concerns traffic safety and the risks of incidents as assumed in figure 2. As road-infrastructure has a serviceable function for motorized mobility a sufficient capacity-management to master the uprush of vehicles nearly all the time is the principal goal of planning and operating the road network. Therefore measures to optimize motorized traffic need a certain balancing between driver’s comfort, traffic safety and prevention of jams as general road traffic regulations are emphasizing. Automatization of road traffic H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 4 Figure 2: Traffic-flow relevant factors of interaction-scenes as a comprehensive figure Criteria weight of traffic relevant factors constituting complexity of changeable frame conditions for interactions on the roadway Degree of Complexity maximal high Density of vehicles on the roadway Velocity of traffic flow normal Degree of freedom in resp. of driving behavior Variety of how acting in cases of interactions Risks for accidents low minimal ( Doerr 2021) A B C D E F Quality of traffic flow (in Level of Service) According to a constant mix of vehicles corresponding to the time window of traffic flow along a phenotypical roadway edge. These frame conditions are fixing the infrastructure capacity usage of the roadway. That theoretical capacity of the roadway is given through its technical features (number of lanes, topographical parameters). • Degree of freedom in resp. of driving behavior: High speed driving while likely none of “opponents” (other vehicles) are adjacent High speed driving is possible but contiguity of opponents has to be observed Modulated velocity is required then some vehicles are adjacent Slow velocity caused by traffic jam • Risks for accidents (from individual caused risks to collective deployed risky situations): Accident caused by own fault (high-risky driving) without any opponent involved Risks caused by misbehaviour within the vehicle´s bulk. A hump in risks is probably because of increasing density of cars while distances between them are diminishing Decreasing risks of heavy accidents while velocity and lane changing are diminishing • Degree of Complexity (for standardizing traffic scenes and predicting steering options) Action-field of high uncertainty predicting traffic interactions because unforeseeable eventualities may occur. Transition from free traffic-flow to bound traffic-flow takes place, so standardized scenes can be observed but the likelihood of incidences increases. That could be a crucial point in respect of “Human-Machine-Interaction” if the artificial “Controlling Master” would be overtaxed. High risky field of incidents whilst velocity of traffic flow is still considerable and the variety of traffic-conditions for steering a vehicle are manifold. High risky field of incidents whilst velocity of traffic flow is still considerable and the Figure 2: Traffic-flow relevant factors of interaction-scenes as a comprehensive figure Traffic planning SCIENCE & RESEARCH International Transportation | Collection 2021 53 operation should follow not only taking individual advantages as manufactures are promoting their cars but also fostering common utility benefits. Besides those expectations road traffic operation could contribute to mitigate problematic effects of motorization aiming at regional environmental and global climate relief. Perception of surroundings for trajectory planning First of all tasks of perception are freed from the question if a car is controlled by a human driver or is partly resp. fully autonomous moving on its pathway. Automated driving of a car needs sensors as detection tools for trajectory planning. For that different sensor technologies installed in a highly automated vehicle - in general four of them - are operating to detect the surroundings 7 . They have distinctive capabilities to perceive static objects and moving subjects in respect of range and resolution. Therefore lots of different detected signals are delivered which have to be subsumed to get a sufficient result as data-base for controlling decisions. One call it “data fusion” playing down the crucial role for the reliability of automated driving cars and the reliance in automatization of road traffic. But a compensation of insufficiencies of each of the sensor techniques to ensure the quality of perception is claiming priority to the end of frictionless trajectory planning. Structuring the surroundings as an analysing scheme as depicted in figure 3 helps to support car-inherent trajectory planning according to the need getting specific data about opponents on the driveway, obstacles along the roadway and places of intervening events which could emerge suddenly. All that information flows into a complex process chain of detection, recognition, interpretation, evaluation, prediction, options generating and an option selecting procedure for claiming an optimum trajectory. To that it belongs a ladder of processing steps within the automatic chain working as cruise control master: Detected signals have to be prepared as data but useless data have to be filtered out of usage. Relevant data has to be interpreted as images. They build the basis for evaluation and prediction of the next seconds of driving in the way generating car-internal scenarios. Prediction deals with probable as well as possible behavior of adjacent road user as interacting players. Therefore that requires a data base in the background of the automaticchain which delivers information about the movingpotentials of the “opponents” to cope with possible risks of interacting with them. The decision-making program has connect that with physical facts along the roadway to make the right decision to occupy the best trajectory, for instance providing an overtaking maneuver. Such scenarios offer certain options for claiming trajectory. Therefore a decisive algorithm has to select the optimum trajectory in order to give steering commands to the machinery of the vehicle. But the decision making software has as background a genuine conditioning program which follows different use cases and realizes mark-typical driving styles according to the marketing of the automobile manufacturer, for instance more sportive for certain premium passenger cars (compare figure 1). The cruise controlling master can be imagined as a chain of a multitude of functions which uses data gained by self-detecting sensors or by external perceived signals via highly efficient telecommunication or “near field” data transfers. The latter serves as data exchange for interconnected operations like vehicle(s)-to-vehicle(s) (v2v) or roadway-infrastructure-to-vehicle(s) (v2i) and vice versa. All these data-inputs have to be processed steadily. But not at least it would be necessary to eliminate useless (in respect of data protection of others) or irrelevant (in respect of trajectory planning) data which can confuse the automat-chain. Otherwise essential data for the control master have to be provided and evaluated which are crucial to construct trajectory-scenarios in respect of options and risks. To this end fields of visions as shown in figure 3 are identified. It seems to be “a little” more complicated than graphic designers are depicting normally autonomous car driving. Composing scenes at phenotypical sceneries To direct the view towards a single vehicle on his pathway or a bulk of cars moving forward should not lead to a limited awareness and incomplete understanding of interactions due to variable traffic conditions. Composing scenes and constructing scenarios at phenotypical structured sceneries might be helpful to arrange realistic tests. The land-use of an urban vicinage along a roadway edge holds either a multitude of intervening interactions spaces as depicted in figure 4. Otherwise the roadway is an isolated trace delimited by parallel barriers like noise protection walls or emergency shoulders as depicted in figures 5-7. Looking at the suburban scenery in figure 4 it represents the empiric approach of a video-drive and delivers heuristic interpretation of intervening interactions as testing features. In contrary of the timed tact of the traffic light regulated intersection in the center a car would be confronted passing these traffic spaces with spontaneous bilateral or multilateral interactions which are hardly to foresee and could provoke an emergency braking or an evasive maneuver. Blind spots which could not be efficient detected due to differences in altitude like hidden doorways of underground garages or ramps of an approaching road should be perceived. But also detection confusing interferences could outgo from the vicinage like dazzling or reflecting by glassy facades of premises nearby. In cases of malfunctions of the carinherent detection the usage of Geographical Information Systems (GIS) feeding the automat-chain of the car H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 6 The cruise controlling master can be imagined as a chain of a multitude of functions which uses data gained by self-detecting sensors or by external perceived signals via highly efficient telecommunication or “near field” data transfers. The latter serves as data exchange for interconnected operations like vehicle(s)-to-vehicle(s) (v2v) or roadway-infrastructure-to-vehicle(s) (v2i) and vice versa. All these data-inputs have to be processed steadily. But not at least it would be necessary to eliminate useless (in respect of data protection of others) or irrelevant (in respect of trajectory planning) data which can confuse the automat-chain. Otherwise essential data for the control master have to be provided and evaluated which are crucial to construct trajectory-scenarios in respect of options and risks. To this end fields of visions as shown in Figure 3 are identified. It seems to be “a little” more complicated than graphic designers are depicting normally autonomous car driving. Figure 3: Fields of vision detected by a solitary car to claim a trajectory onto his pathway Composing scenes at phenotypical sceneries To direct the view towards a single vehicle on his pathway or a bulk of cars moving forward should not lead to a limited awareness and incomplete understanding of interactions due to variable traffic conditions. Composing scenes and constructing scenarios at phenotypical structured sceneries might be helpful to arrange realistic tests. The land-use of an urban vicinage along a roadway edge holds either a multitude of intervening interactions spaces as depicted in Figure 4. Otherwise the roadway is an isolated trace delimited by parallel barriers like noise protection walls or emergency shoulders as depicted in Figure 5-7. Looking at the suburban scenery in Figure 4 it represents the empiric approach of a video-drive and delivers heuristic interpretation of intervening interactions as testing features. In contrary of the timed tact of the traffic light regulated intersection in the center a car would be confronted passing these traffic spaces with spontaneous bilateral or multilateral interactions which are hardly to foresee and could provoke an emergency braking or an evasive maneuver. Blind spots which could not be efficient detected due to differences in altitude like hidden doorways of underground garages or ramps of an approaching road should be perceived. But also detection confusing interferences could outgo from the vicinage like dazzling or reflecting by glassy facades of premises nearby. In cases of malfunctions of the car-inherent detection the usage of Geographical Information Systems (GIS) feeding the automat-chain of the car with physical data about critical intervening points along a road can serve as warning resp. corrective instrument. Apart from V2V-communication between adjacent vehicles exogenous data-imports via Global Positioning Systems (GPS) need either a field of open visions or road users equipped with GPS-communication-tools as transponder. As local equipment of the roadinfrastructure wire lines integrated in the road surface or transponders mounted on pylons overhead could installed if the road provider will be willingly to bear the costs for it viii . S a f e t y b u b b l e K e e p c l e a r l e n t i l S a f e t y b u b b l e T r a j e c t o r y s p a c e S u r r o u n d i n g s R o a d w a y U r b a n / r u r a l l a n d - u s e v i c i n a g e S h o u l d e r S c e n e r y P a t h w a y e d g e (I n t e r a c t i o n ´ s b o x) F r o n t s i d e d e t e c t i o n t ion E m e r g e n c y l a n e S p e e d l a n e T r a n s i t l a n e E x i t l a n e Proprietary land-use / built-up front line Public urban space / distance green R o a d w a y´s u r b a n s p a c e Figure 3: Fields of vision detected by a solitary car to claim a trajectory onto his pathway SCIENCE & RESEARCH Traffic planning International Transportation | Collection 2021 54 with physical data about critical intervening points along a road can serve as warning resp. corrective instrument. Apart from V2V-communication between adjacent vehicles exogenous data-imports via Global Positioning Systems (GPS) need either a field of open visions or road users equipped with GPS-communication-tools as transponder. As local equipment of the road-infrastructure wire lines integrated in the road surface or transponders mounted on pylons overhead could installed if the road provider will be willingly to bear the costs for it 8 . Boundary conditions for constructing scenarios If the compositions of scenes have cleared up the boundary conditions for constructing scenarios in the mentioned way prospective assumptions of future road traffic events could be put up. For this “Real World” forms the starting point to attribute future technological features to the vehicles like SAE-Levels and to assume the usage of them if a driver could select a driving mode like downgrading the automated functions (figures 7 and 8). But not A section of Danube-Embankment-Highway (A22) in Vienna as basic scenery Composing a field of scenes as can be observed nowadays the action field (“playing ground”) is defined by a plausible delimited (entry-exit) roadway edge. The observation of scenes begins with a starting shot (t 0 ) and terminates if the last of the players has left the edge passing an exit line (t n ). So a coincident collective of cars could be proven in respect of open trajectory options for moving forward. It forms the base for constructing scenarios assuming a variety of automated vehicles on different levels under varied driving modes in respect of driver´s presence. A section of Danube-Embankment-Highway (A22) in Vienna as basic scenery Constructing a field of scenario A: Hereby, it is assumed that all vehicles are under remote control of the traffic management of the highway section. Therefore all cars have to be exclusively equipped for external interconnectivity “vehicle to infrastructure” via 5+G-communication and for remote steering. Car-drivers have to accept the rule being out of steering their vehicles. Such a traffic operation results in a homogeneous traffic flow ensuring distance keeping and velocity control. But the driver has to declare his trip plan in advance especially which exits should be frequented. A section of Danube-Embankment-Highway (A22) in Vienna as basic scenery Constructing a field of scenario B: This scenario is based on the general assumption that each vehicle has its own specific characteristic in respect of motor power and automat-level. It seems to be the most realistic as well as the most complex scenario of partly automated road traffic under the principle of an indiscriminate traffic participation of motorists far-off. Each vehicle behaves itself as a solitary actor. Such a mixture of cars causes a heterogeneous nearly unpredictable flow of traffic. Therefore the prevention of critical interactions will be focused on. Figure 5-7: Observing scenes and constructing scenarios of automated traffic on a pathway edge. Left to right: Section of Danube-Embankment- Highway (A22) in Vienna as basic scenery. Section control there as test pathway. Ubiquitous phenotypical section seen in Berlin (A100 Exit Halensee) H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 7 Figure 4: Intervened interactions spaces as phenotypical characteristic of a radial urban trunk road Outlined by Viktoria Marsch and Andreas Romstorfer 2017. Scenery: Intersection Bruenner Road/ Katsushika Street in Vienna-Floridsdorf Boundary conditions for constructing scenarios If the compositions of scenes have cleared up the boundary conditions for constructing scenarios in the mentioned way prospective assumptions of future road traffic events could be put up. For this “Real World” forms the starting point to attribute future technological features to the vehicles like SAE-Levels and to assume the usage of them if a driver could select a driving mode like downgrading the automated functions (Fig. 7 & 8). But not only drivers are actors. Road network providers could play a dominating role to manage the traffic via steering each vehicle in designated section of their network assisted by telecommunication providers. In this respect different scenarios could be drafted aiming at cutting traffic rush or reducing inflow of vehicles to prevent jams. Used cars dealer Furniture shopping center Gas station Car dealer Entry/ exit Exit Entry Entry/ exit Entry/ exit Car dealer Exit Gas station Pedestrian’s sidewalk Bicycle lane Lanes for motor cars Rolling board for advertisements Intervening car maneuvers Key for Interactions space Gas station Exit Access for pickup commodities Residential block with kindergarten Commercial parking lot Entry/ exit Entry/ exit Key for Area of traffic phenomena Special lane for cyclists Parking lane Narrowing of carriageway Spreading of carriageway (additional lane) Right turning car yielding at an exit Right turning car entering in a proprietary site Entry/ exit underground Parking Key for Interactions space Sidewalk Parking lane to pick up kids Lane for car traffic flow Street car stop platform Bus stop point Driving manoeuvres Spontaneous crossings of pedestrians Figure 4: Intervened interactions spaces as phenotypical characteristic of a radial urban trunk road Outlined by Viktoria Marsch and Andreas Romstorfer 2017. Scenery: Intersection Bruenner Road/ Katsushika Street in Vienna-Floridsdorf Traffic planning SCIENCE & RESEARCH International Transportation | Collection 2021 55 only drivers are actors. Road network providers could play a dominating role to manage the traffic via steering each vehicle in designated section of their network assisted by telecommunication providers. In this respect different scenarios could be drafted aiming at cutting traffic rush or reducing inflow of vehicles to prevent jams. A systematic monitoring of scenes seen from an appropriate observation point as here a S-Bahn-platform (figure 9) delivers a lot of interactions which can be standardized and statistical documented. But also eventualities might occur as result of irregularly and risky driving-maneuvers. It gives deductive ideas for testarrangements which could be simulated in realistic ways on testing grounds. Test and Implementation Procedure as result From the discussed issues an interdisciplinary structured and phased test and implementation procedure (“TIP”) is resulted. It will integrate all stakeholders of the automotive manufactures as well as all concerned groups of the mobility system. The test procedure depicted in figure 10 describes a chain which has its beginning at testing grounds within the responsibility of the automobile sector and is terminating in the Real World of our habitat. So one could realize points of transition within the succession of phases where the concerns are changing between the productive automobile sector, the approving authorities, the bearer of infrastructure, the driver’s instructing schools, the interest-groups of mobility participants included motorists, the planning community and other more. Acting stakeholders and involved professionals affiliated to the automotive industrial sector, in particular the testing ground staff, the test designing programers, the on-board testing team and analysts of results, are participating in a bound testing program managed by an automotive OEM and connected suppliers. In its core testing components are proven in respect of functionality and reliability. If the test results let guess to be sufficient enough for deployment in road traffic legal authorities have to be concerned for authorization. In this way deployment on the automobile market can start accompanied by promotion to awaken car-buyer’s awareness and to influence public opinion. Media are reflecting willingly the marketing messages. Foremost after that it could be a topic within the frame of the New Cars Assessment Program (NCAP) of national Driver’s Associations to strengthen or disprove the effects of automatic functions resp. driving assistance systems. In the next phase policies will be possibly challenged: On the transnational level the United Nations Economic Commission for Europe (UNECE) has founded a working party for that, recently to harmonize rules for deploying cars on level 3. On the national level the legislation has to adapt general driving rules, especially in respect of the handling of semi-autonomous driving onto the road network, what concerns so-called Human-Machine-Interaction. But as well on a regional or local level competent authorities could have been exerted pressure by citizens to direct local regulations. As figure 10 indicates each step has a specific personell-configuration according to the tasks to do and the qualification to do it. Some steps are inevitable resp. obligatory to carry out (drawn arrows) others are successively forced (dashed arrows) or at least advisably (dotted arrows) to be considered. Furthermore it is readable who plays the initiative part as stakeholder-group and who is acting in practice as a team. Such details to discuss is reserved for a long version. Foresighted public entities providing road infrastructure and being responsible for public spaces might integrate such critical issues in their development plans. Also driving school’s instruction should receive supplementary contents and could offer training about it. By the way this group of essential concerned professionals is seldom consulted. Résumé for an open discourse Approaches derived from Real World as exemplified do not solve technical problems as that are tasks for auto- H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 9 Figure 8: Construction of a scenario field onto a three-lane carriageway covered by a mixture of vehicles representing different levels of automation and vehicle classes moving in a solitary mode A systematic monitoring of scenes seen from an appropriate observation point as here a S-Bahn-platform (Fig. 9) delivers a lot of interactions which can be standardized and statistical documented. But also eventualities might occur as result of irregularly and risky driving-maneuvers. It gives deductive ideas for test-arrangements which could be simulated in realistic ways on testing grounds. derived from the “Real World” of phenotypical sceneries. Figure 9: Observation of scenes at an access to an urban ring-highway with continuous traffic flow during the day (3s-timing). Scenery: Berlin A 100 (“Hundekopf”) Messedamm (at a working day before noon) Critical interactions when converging Distance keeping as critical driving maneuver Sufficient distance keeping Not sufficient braking distance Non regular converging over blocking line Motor cyclists changing lane to prevent incidents Figure 9: Observation of scenes at an access to an urban ring-highway with continuous traffic flow during the day (3s-timing). Scenery: Berlin A 100 (“Hundekopf”) Messedamm (at a working day, before noon, in September) Figure 8: Construction of a scenario field onto a three-lane carriageway covered by a mixture of vehicles representing different levels of automation and vehicle classes moving in a solitary mode H. Doerr et al.: Assessment of autonomous moving cars: Approaches and test procedures 9 Figure 8: Construction of a scenario field onto a three-lane carriageway covered by a mixture of vehicles representing different levels of automation and vehicle classes moving in a solitary mode A systematic monitoring of scenes seen from an appropriate observation point as here a S-Bahn-platform (Fig. 9) delivers a lot of interactions which can be standardized and statistical documented. But also eventualities might occur as result of irregularly and risky driving-maneuvers. It gives deductive ideas for test-arrangements which could be simulated in realistic ways on testing grounds. derived from the “Real World” of phenotypical sceneries. Figure 9: Observation of scenes at an access to an urban ring-highway with continuous traffic flow during the day (3s-timing). Scenery: Berlin A 100 (“Hundekopf”) Messedamm (at a working day before noon) Critical interactions when converging Distance keeping as critical driving maneuver Sufficient distance keeping Not sufficient braking distance Non regular converging over blocking line Motor cyclists changing lane to prevent incidents Figure 9: Observation of scenes at an access to an urban ring-highway with continuous traffic flow during the day (3s-timing). Scenery: Berlin A 100 (“Hundekopf”) Messedamm (at a working day, before noon, in September) Figure 9: Observation of scenes at an access to an urban ring-highway with continuous traffic flow during the day (3s-timing). Scenery: Berlin A 100 (“Hundekopf”) Messedamm (at a working day, before noon, in September) SCIENCE & RESEARCH Traffic planning International Transportation | Collection 2021 56 motive research and development. Rather it should help to trace out challenging traffic events and to put up framing conditions which influence traffic flows exogenously. In such a multidisciplinary manner deficiencies untied from pure technical quality requirements and standardization, like ISO 26262, could be revealed 9 . Embedded in a chain of test and implementation procedures all relevant stakeholder and affected groups from the mobility milieu can be addressed. Not at least, because this methodical approaching should enable them to reflect the evoked changes by the arising innovations within the mobility system and to encourage them to contribute their considerations to that in a democratic discourse. ■ 1 As there are Advanced Driver Assistance Systems which equip vehicles on SAE-Level 3 (or more) helping driving maneuvers like distance keeping, lane keeping, speed limitation, overtaking other cars or parking-maneuvers. Furthermore driving on a higher level as Autonomous Driving System will allow a transition to partly or fully autonomous movements of the car onto designated section of the public road network. In the opinion of the author this stage is envisaged not before the thirties or forties of our century. A sceptic perspective to it delivers a study by the ADAC e.V. (2018): ADAC e.V. (2018): Einführung von Automatisierungsfunktionen in der Pkw-Flotte. Auswirkungen auf Bestand und Sicherheit. Erstellt von Prognos GmbH. Berlin / München 2 Such scenes have been exemplified by figure 2 and figure 4 depicted in: H.-Doerr, A. Romstorfer: Implementation of autonomous vehicle onto roadways. Internationales Verkehrswesen (72) 1/ 2020, 67, pp 66-70; and in: International Transportation | Collection 2020, pp 58-62 3 To this end a comprehensive “vademecum” (a kind of informal manual) has been prepared, which contains methodical help to cope with the complexity of the topic in a multidisciplinary view. Advices are given in form of an alphabetic ordered terminology which is practical oriented to address practitioners. It is not published yet as a whole, but the German version would be ready. Some figures depicted here are taken from it. 4 Look at the endnote 2 5 Five levels are usually distinguished according to: Society of Automotive Engineers International, 2014 SAE J3016: Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems. http: / / standards.sae.org/ j3016_201401/ 6 The Level-of-Service-Indication A-E is derived from the US Highway Capacity Manual (2000), which deals with connexions of urban street classes and velocity of traffic flow in respect of capacity criteria and driving strategy. 7 As there are: Cameras as hardware plus object-recognition and image-interpretation as software, Light-Detection and Ranging based on Laser-technology (LiDAR), RADAR as well-known for area-detection, Ultrasound-detection very close to the car and other means more (e.g. infrared). 8 Compare with figure 2 depicted in Internationales Verkehrswesen/ International Transportation 1/ 2020, p 68. 9 This quality management terms of reference to prove electronical instruments are mostly used in respect of technical reliability of automatic functionalities. But this limited procedure does not fulfil sufficiently requirements of an assessment in respect of traffic aspects. Andreas Romstorfer, Dipl.-Ing. (FH), MA arp-planning.consulting.research, Vienna (AT) a.romstorfer@arp.co.at Viktoria Marsch, DDipl.-Ing. Traffic researcher and planner at arp 2013-2017, Vienna (AT) viktoria.marsch@aon.at Heinz Dörr, Dr. Dipl.-Ing. Consulting engineer spatial and traffic planning, arp-planning.consulting.research, Vienna (AT) heinz.doerr@arp.co.at DRAFT OF AN INTERDISCIPLINARY STRATEGIC TEST-PROCEDURE FOR THE AUTOMATIZATION AND AUTONOMIZATION OF VEHICLES ONTO ROADWAYS General Steps phased Phase of Test-Conception Phase of Test-Implementation Phase of Test-Programing Phase of Test-Realization Phase of Test-Result-Evaluation Phase of legal Authorization Phase of Deployment Phase of Traffic Policies Phase of Verification Phase of Adaption Phase of Implementation Restructuring built-up areas Changing of mobility patterns Knowledge-Transfer to related professionals Driving instruction & training Consideration in Roadway Design Addition of traffic statistics Adaption of traffic rules Observing interaction scenes Surveying traffic behavior Monitoring of incidencies Civic inquiry Integration into Mobility Action Plan Local traffic regulations Roadnet organisation Equipped cars emerging onto roadways Public awareness & reflection Product placement / migration on markets Legalization of technological components Proving legal conformity Validation of functionality Lay open the testing result Test-Goals Test-Vicinage Testing Ground Testing-formation Testing-elements Test Purpose Test Objects Evaluation of traffic capability Identification of Insufficiences Test-Recording Stage the interaction-scenes Provision for resources Script of Scenarios Setting the theatre Testing Tasks Figure 10: Test procedure as a phased chain divided in automotive testing arrangements, legal authorization, deployment on vehicle’s stock and final implementation in road net and mobility system