WIXLIB

Background: What are vehicles with self-driving mode?Next chapterHumans continue to takeover responsibility in thesystemAlthough self-driving mode can replace the driverafter its activation, it cannot replace all humans inthe system.If there is no driver for a vehicle, humans in other functionstake on some of the responsibility instead. In recentresearch projects – and everyday language – vehiclesreferred to as “autonomous” often have a safety driverwho intervenes in critical situations. In these vehicles,self-driving mode is still under development. Theconsequence of this is that the driver of such vehicleshas more responsibility than a conventional driver. Theauthorities that mandate the piloting companies requiresuch vehicles to have a fully attentive licensed driverwith special training behind the wheel at all times. Forexample, when issuing a “Testing Permit,” the New YorkState Department of Motor Vehicles states that “A personholding a valid driver license must be in the driver’s seatwhile a vehicle is operated on public highways.15E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterThat person must be prepared to take control whenrequired to in order to operate the vehicle safely andlawfully. Every test vehicle operator must be adequatelytrained in the safe operation of the test vehicle to ensureboth legal and safe operation.”2 In the US, there is novehicle currently available for sale that is self-driving.3Even if in the future, with no drivers in the vehicle and theauthorities not actively requiring one, there will be somesort of supervising vehicle attendant inside or outsideof the vehicle. At the same time, the monitoring andmaintenance staff at depots and in operating centers(or control centers) can play an important role. For along time to come, those responsible – either insideor outside the vehicle – must be available to performsafety checks on the vehicle before start-up, to activateand deactivate the self-driving mode, and to guaranteereliable functioning while it’s in operation. In exceptionalcases, it may even be necessary for an “emergencydriver” to take over operation on-site or remotely.Furthermore, the authorities expect the applicant fora vehicle permit to “submit with this application a lawenforcement interaction plan to inform law enforcementofficers and first responders how to safely interact withthe [ ] vehicle(s) in emergency and traffic enforcementsituations.”416E-book: Self-driving and cooperative cars

KapitelnameThe technologies behindthe autonomyFor a vehicle to move without driver intervention, it mustitself have a sufficient wealth of information. In orderto obtain this information and to provide this relevantinformation to the vehicle, various technical developmentsoutside the vehicle are required. The following table givesan overview of the technology that enables self-drivingmode, making it easier to understand what these vehiclescan do.17E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterInformationTechnologiesDriving order: The A human always determines the driving order. Thevehicle must “know”vehicle cannot decide “autonomously” where itwhich destination it isis going, the human must input the destinationexpected to drive to.themselves in some form or another (e.g. “Drive to thenext charging station”). The digitalization of communication can help here, forexample, apps and smartphones.Self-positioning: TheMost of today’s transport positioning systems use thevehicle must “know”satellite-based Global Positioning System (GPS).where it’s located.GPS guarantees accuracy of positioning with deviationsbelow eight meters. For self-driving mode, deviations of lessthan 5 cm may be required. This can be guaranteed by: Differential Global Positioning System (DGPS):Reference stations on the earth eliminate GPSinaccuracies. Landmark and marker orientation: If the systemuses its sensors (camera) to recognize prominentlandmarks whose position is stored in the system; theycan position themselves with them. Road markings(center line) are also gathered to aid orientation.18E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterInformationTechnologiesSelf-positioning: The “Virtual Rail” with Radio Frequencyvehicle must “know”Identification (RFID) or other processes: Smallwhere it’s located.chips or other electronic orientation aids are appliedin or on the road or in the road environment, whichcorresponding transmitter-receiver systems can detectelectromagnetically to use for self-positioning. If the speed and direction of travel of the vehicle areknown due to vehicle data, a position change can becalculated.Self-status detection:Many different sensors provide this information:The vehicle must know aOdometers, rotation rate sensors on the wheels,lot about its own status:gyrocompass, tire pressure sensors, door openingFunctioning of importantsensors, thermometer, etc.systems, direction oftravel, speed, door closing,indoor and outdoortemperature, etc.19E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterInformationTechnologiesRoadway and envi-From static geodata (map data), to static traffic elementsronmental conditions:(traffic signs, construction sites), to traffic light signals, theThe vehicle must “know”position of fog banks, traffic congestion, individual cars,how to reach its desti-bicycles, and pedestrians, a motor vehicle with self-drivingnation the fastest andmode must have an accurate picture of its surroundings.safest way.A “Local Dynamic Map” (LDM) stores the requiredTraffic situation: Thedata in the vehicle. Static information (e.g. map data) canvehicle needs the exactbe stored in the LDM in advance and compared with theposition of buildings andactual environment in test runs on the intended route.other fixed obstacles,However, data fed in this way requires constant updating.the current position, theThe LDM may obtain data from sensors on the vehicleoutline, the speed, anditself or from sensors or digital services outside the vehicle.the direction of move-( Illustration).ment of all road users inthe area (cars, bicycles,pedestrians), traffic signs,traffic lights, etc. Onlywith a precise picture ofthe traffic situation canit also take into accountall traffic regulations andavoid the endangermentof other road users. To satisfy the high demand for data, the vehicle mustbe networked with its environment. The followingserve as communication: Mobile communications: The large amountsof data can be handled better with a high mobilestandard. The currently much-discussed 5G mobilestandard seems particularly suitable here given itshigh data transmission rates, real-time transmission,and low latency. Wi-Fi: A wireless local area network (Wi-Fi) is beingbuilt by wireless modules on streetlights and signgantries to improve communication.20E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterInformationTechnologiesRoadway and envi- Sensors in the vehicle or on the roadside detect theronmental conditions:The vehicle must “know”how to reach its destination the fastest andsafest way.Traffic situation: Thevehicle needs the exactposition of buildings andother fixed obstacles,the current position, theoutline, the speed, andthe direction of movement of all road users inthe area (cars, bicycles,pedestrians), traffic signs,traffic lights, etc. Onlywith a precise picture ofthe traffic situation canit also take into accountall traffic regulations andavoid the endangermentof other road users.21traffic environment and situation: Digital camera: Cameras capture the environmentoptically, but may not detect everything (e.g. indarkness, fog, or backlight). Laser scanner/LIDAR: The distance and speedof objects can be measured by using many laserbeams. This allows for the creation of an accuratepicture of the environment. Radar and ultrasound sensors: Additionaldirectional and distance measurement technologiescomplement the range of perception of othersensors and provide redundancy so that nothing isoverlooked. Sensor data consolidation and evaluation:Different sensors and the images they have createdand the data they collected have various advantagesand disadvantages. Computers can use it to calculatea coherent picture that detects traffic much better thanjust one sensor. Regional digital map: Geodata and dynamictraffic data can be merged and refined in a vehicleindependent digital map. This data can be used by thevehicle.E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterInformationTechnologiesRoadway and envi- Sensor data from other vehicles (Vehicle-to-VehicleCommunication, V2V) can also be used, for exampleto supplement a regional digital map (e.g. trafficcongestion information from other vehicles).ronmental conditions:The vehicle must “know”how to reach its destination the fastest andsafest way.Traffic situation: Thevehicle needs the exactposition of buildings andother fixed obstacles,the current position, theoutline, the speed, andthe direction of move- Digital traffic signs: A camera cannot detect withsufficient reliability whether a traffic light is green or red.A digital radio signal from the traffic light can provideclarity in this case. Control technology: Brings together a large amountof data contained in an operator center in order tobe able to centrally monitor and maintain the vehicleand infrastructure and coordinate the necessaryinterventions from outside.ment of all road users inthe area (cars, bicycles,pedestrians), traffic signs,traffic lights, etc. Onlywith a precise picture ofthe traffic situation canit also take into accountall traffic regulations andavoid the endangermentof other road users.22E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterExample of a “Local Dynamic Map”VehiclePedestriansEgo vehicleType 4Slippery roadSignal phaseType 3Traffic congestionLandmarksType 2Traffic signMapType 1Type 1: Permanent static data (map data)Type 2: Transistent static data (roadside infrastructure)Type 3: Transient dynamic data (congestion, signal phase)Type 4: Highly dinamic data (vehicles pedestrians)23E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterPowerful computer programs and artificial intelligenceIn particular, the merging and evaluation of sensor data requires the use of powerfulcomputers and complex programs. In this case, artificial intelligence can also be used.What does artificial intelligence mean?Artificial intelligence describes the ability of acomputer program to extract logical correlations fromlarge amounts of information (pattern analysis andpattern recognition) to draw conclusions according tocertain rules (i.e. predicting patterns).These processes are often used in image recognition and image analysis. For example,if a computer program has already analyzed tens of thousands of images of differentmodels of cars, it will be able to identify a car as such even if it has not “seen” the specificmodel before. Accordingly, if the computer program has analyzed many vehicles turns,it can predict the course of a turning vehicle with certain probabilities. This ability is veryimportant for self-driving mode since the systems are better able to navigate in unknownsurroundings and constellations. Independent and ethically thought-out decisions,comparable to those of a human being, cannot be made by such systems. Therefore, theterm “artificial intelligence” may give rise to partially exaggerated expectations.24E-book: Self-driving and cooperative cars

KapitelnameCooperative driving: From vehicle systemto integrated overall system?The self-driving vehicle system of a car with self-driving mode may, in certaincircumstances, rely on digital surrounding environment systems (cooperative automation).5From a technical point of view, the vehicle system with digital surrounding environmentsystems can also be considered as an integrated overall system. This overall systemmust meet certain requirements to ensure optimum results for traffic safety and trafficflow. Autonomous driving (SAE: cooperative driving) requires a change of focus from thevehicle to an overall system because it will only achieve adequate performance in varioustraffic situations in close cooperation with digital surrounding environment systems. Forexample, when driving through traffic lights at a complicated intersection on-road sensorsand digital signals from the traffic lights themselves would be needed at a red light toensure safety and accuracy.We are accustomed to associating cars almost as a uniform or a distinct tool used by adriver. This distinction is common to all traffic and vehicle safety regulations in every country.In the US the National Traffic and Motor Vehicle Safety Act6 regulates motor vehicle safetyand mandates the application of the Federal motor vehicle standards7 regarding design,construction and performance of all motor vehicles.825E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterThen again the rules of the road are regulated by the Federal States and address the driveronly, who must control his tool and is responsible for every action of the vehicle itself. Ifthe driver in the vehicle becomes just a passenger, this division can no longer work. Carmanufacturers must comply with Federal Motor Vehicle Safety Standards and certify thattheir vehicle well aligned with safety features to avoid risks under all circumstances and inall driving modes. For motor vehicles with autonomous motion functions this includes toensure that they operate in compliance with the rules of the road.9The concept of the vehicle as a definable tool must also be questioned. Whether a vehiclewith self-driving mode can handle its driving task safely and smoothly will increasinglydepend on information originating from various digital surrounding environment systems(digital traffic signs, regional digital map, roadside sensors, etc.) (cooperative drivingsystems). The vehicle drives based upon its own digital map, which it continuously updateswith the help of information from outside. The vehicle is not always able to decide on itsown whether this data is reliable or not. It has to rely in part on the surrounding environmentsystem. The U.S. Department of Transportation states that cooperative automation willbe “an important complementary technology that is expected to enhance the benefits ofautomation at all levels, but should not be and realistically cannot be a precondition to thedeployment of automated vehicles.”10“I think that SAE Level 5 will remain adream for quite a long time, but SAE Level4 Technology V2V and V2X will be verybeneficial and more important than whatthe current discussion implies for themajority of valuable use cases.”Matthias Hartwig26E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterTraffic light example A vehicle cannot detect a conventional traffic light with acamera with enough reliability. It can only drive quickly at an intersection if it gets a digitaltraffic signal. However, this signal only has an added value if the trafficlight operator guarantees its reliability. The vehicle is thus increasingly cooperative with digitalsurrounding environment systems.Parking garage example A parking garage designed for self-driving valet parkingcould soon be possible using comprehensive data wherecameras and other sensors detect free parking spaces,vehicles, and pedestrians into a dynamic traffic map(technical surrounding environment system). Navigating becomes very easy for the motor vehicle inself-driving mode here. However, it must be able to relyon the given information. Not only does the vehicle haveto be safe, but the vehicle and surrounding environmentsystem must form a secure overall system to preventaccidents together. At a certain point, the question arises as to who is actuallyin control of the vehicle – the vehicle or the parkinggarage.27E-book: Self-driving and cooperative cars

Background: What are vehicles with self-driving mode?Next chapterThe vehicle and the surrounding environment system would technically form an overallsystem. Only if this overall system is safe and provides enough valid data for navigation willthe vehicle be able to safely navigate and run smoothly through traffic. In order to ensurethe identifiability of those responsible, there must be continual legal areas of responsibility.As a result, in the future it is likely that there will be system administrators for the vehiclewhom will monitor the various digital surrounding environment systems. Technically, asingle system in isolation would often lead to a dead end as the required amount of safetyprecautions along with the monitoring of traffic flow can only be provided when there isan interaction of all component

3 E-book: Self-driving and cooperative cars The Author Next chapter Assessor jur. Matthias Hartwig is a fully qualified lawyer and has been a scientific advisor to the Institute for Climate Protection, Energy and Mobility (IKEM) since 2011. Leading the Mobility Team since 2014, he also deals with the legal issues surrounding mobility.