Sensor Systems and Communication Technologies in Autonomous Driving

Table of Contents

1 Introduction

2 General Facts of Autonomous Driving
2.1 Definition
2.2 Classification
2.3 Functionality
2.4 Potentials and Challenges
2.4.1 Social Factors
2.4.2 Ecological Factors
2.4.3 Economic Factors
2.4.4 Challenges
2.5 State of the Art of Autonomous Driving

3 Technical Realisation
3.1 Sensor Systems
3.1.1 Ultrasonic Sound
3.1.2 Cameras
3.1.3 Radar (Radio Detection and Ranging)
3.1.4 Infrared
3.1.5 LIDAR (Light Detection and Ranging)
3.1.6 Comparison of Different Sensor Systems
3.2 Vehicle Control Unit
3.3 Communication Systems in the Vehicle
3.3.1 Classification
3.3.2 Composition
3.3.3 ISO 9141-K-Line
3.3.4 LIN (Local Interconnect Network)
3.3.5 CAN (Controller Area Network)
3.3.6 FlexRay
3.3.7 MOST

4 ISO 26262
4.1 Application Area
4.2 Content

5 Conclusion

6 Bibliography

7 List of Figures

8 List of Tables

List of Abbreviations

Kurzfassung

Diese Bachelorarbeit gibt einen Überblick über autonome Fahrzeuge und die Funktion innerhalb eines Vehikels. Neben der technischen Umsetzung sind auch Potentiale und Herausforderungen darzulegen. Die Vision des unfallfreien Fahrens steht hinter der Idee der selbststeuernden Fahrzeuge.

Die Assistenzsysteme werden von Jahr zu Jahr mit dem Ziel des autonomen Fahrens verbessert. Es ist hier zwischen mehreren Sensorsystemen (RADAR, LIDAR, Kamera, Infrarot, usw.) zu unterschieden, welche je nach Einsatzgebiet verbaut werden.

Die Anforderungen an das Kommunikationssystem werden mit der steigenden Anzahl an Sensoren und der immer größeren zu übertragenden Informationen im Fahrzeug deutlich erhöht. Es ist hier ebenfalls abzuwägen, welches System für welche Anforderungen eingesetzt werden soll, um die günstigste Alternative festzulegen.

Ziel ist es, dem Leser das nötige Wissen zu vermitteln, was unter einem autonomen Fahrzeug zu verstehen ist, welche verschiedenen Sensorsysteme und welche Kommunikationssysteme es im Fahrzeug gibt, aber auch welche Norm in der Automobilbranche zum Einsatz kommt. Die Informationen werden detailgetreu, aber trotzdem einfach beleuchtet.

Abstract

In this bachelor thesis an overview of autonomous vehicles and the function inside the automobile is given. Besides the technical implementation, potentials and challenges have to be exposed. The vision of accident free driving stands behind the idea of self-controlled vehicles.

Assistant systems will be enhanced year after year with the objective of autonomous driving. Different sensor systems (RADAR, LIDAR, camera, infrared, etc.) have to be distinguished which are obstructed depending on the operation area.

The requirements to the communication system will obviously increase by a rising number of sensors and the information which have to be transferred in the vehicle. The communication systems have to be balanced to decide which is the best and the cheapest one to meet the requirements.

This thesis was written aiming at imparting knowledge to the reader what does an autonomous vehicle mean, which sensor systems and communication systems are there and which norm is deployed in the automobile industry. It is important to interfere the information detailed, but still easy.

Acknowledgements

First of all, I want to thank my family for supporting me at this bachelor thesis. I could ask them every question and got constructive critic to finish my thesis. I also want to thank Dipl.-Ing. Walter Wölfel, who supervised my thesis. Because of him I had the idea to write about Autonomous Driving. He has fully supported me and if needed, I got comprehensive information.

1 Introduction

The automotive industry is in phase of fundamental changes. Although autonomy vehicles were in the past merely objects of science-fiction-films and books, nowadays the question is not if autonomous vehicles are able to be built but when it will finally happen. The technology in this day and age is able to manage specific traffic situations. Audi, BMW, Tesla and Google demonstrated with their prototypes that autonomous vehicles can navigate independently without any input of the driver.

The two most substantial arguments for autonomous driving are increasing the safety on the streets and reducing emissions. 90 percent of accidents happen by the reason of human mistakes. By means of autonomous systems an enormous decrease of accidents will be achieved. Furthermore, due to the optimisation of the traffic flow a reduction of CO2-Emissions will be the impact.

Meanwhile, the broad public recognised the relevance of autonomous driving und supposed themselves to the topic. Adjacent to the increasing news coverage the user acceptance on social lines is in the focal point.

In the literature a high number of papers about autonomous driving exist. Several theses discuss mainly ecological impacts while other analyse the potentials and risks. Technical aspects of autonomous vehicles are superficial given in papers. On this academic void this thesis comes in the present input and informs about the technical implementation detailed.

The different automation classification and the functionality is shown. Also the state of the art, potentials and challenges are examined.

For autonomous systems, a multitude of sensors and measuring devices are used. The function of these sensors will be explained. Advantages and Disadvantages of the systems will be balanced to decide which sensor is the best for an area of operations. Another important point is the communication inside the vehicle. Sensors have to communicate with the vehicle control unit. Different possibilities can be used for a working communication between the sensors and the control unit. The heart of the vehicle is the control unit. Every signal is handled by means of it and the control of the peripheral elements is performed. Enough processing power is used to handle the signals in real-time.

The most important standard for automotive vehicles the ISO 26262 will be explained.

Judicial and ethic aspects will not be discussed in this work.

2 General Facts of Autonomous Driving

Generally, it is important to know what is autonomous driving and how does it works. Potentials and challenges have to get reflected. It is also interesting what is state of the art and what is already developed.

2.1 Definition

The term ‘Autonomous’ comes from the Greek and means ‘one who gives oneself one’s own law’. Historically the word autonomous was used for vehicles, which were moved without pulling or pushing by animals. Since it’s technically feasible to control vehicles without human influence the term automated driving gets new defined.

Today a vehicle is autonomous if it moves itself partly or fully. The driver doesn’t have to do any manoeuvres like steering, breaking or controlling the speed depends on the level of automation. (Knott, 2016)

2.2 Classification

Six different levels categorize the automation of vehicles. These classifications are used in Europe and the U.S.

- Level 0: The driver has to drive on his own. No Assistance Systems are available but warnings can exist.
- Level 1: The driver has to control the vehicle. Assistance Systems like Adaptive Cruise Control or Parking Assistance are available.
- Level 2 (Part automation): The vehicle is able to accelerate, brake and steer itself. The driver has to respond if the automated system fails. The system will also deactivate immediately if the driver takes over the control.
- Level 3 (High Automation): The system runs self-reliant functions like activating the indicator, makes lane changes and holds the lane. The driver can be informed by the system if the attention of the driver is needed. The lawgiver is working to declassify it. Some people say, till 2020 level 3 should be allowed.
- Level 4 (Full Automation): The system drives on its own. If there are some special events the driver could be needed to drive the car.
- Level 5: The Vehicle is able to drive everywhere it is legal to drive. The driver just has to define the target and to start the system. (Knott, 2016)

2.3 Functionality

An autonomous system uses the same information that a human driver needs for steering a car. It has to overview the environment of the car and to process this information for controlling the car. This Information has to be enough for the system to operate the car. By means of several technical sensors the autonomous system is able to realize obstacles or other cars earlier than humans can. These sensors are obstructed on every side of the car to detect the environment also around the vehicle. If a sensor realises an obstacle this information gets to the control unit. The control unit can act and change for example the steering position and the car changes its driving direction. (Rügheimer, 2016)

2.4 Potentials and Challenges

With automated systems potentials are possible but also challenges have to get overcome.

2.4.1 Social Factors

Automated vehicles are 10times less convoluted in an accident than normal vehicles which are controlled by humans. Machines can “see” with several sensors, so the risk of an accident decreases to a very low value. The insurance market will have to change because the insurance will pass from the driver to the automotive manufacturer. (Reiters, 2016)

Another social factor is the time the driver can safe by driving an automated car. The driver does not have to pay attention on the traffic and can spend the time usefully during sitting behind the wheel. It does not matter whether the driver sleeps or works (Figure 1). (Oliver Wyman, 2015)

Abbildung in dieser Leseprobe nicht enthalten

Figure 1: Driver becomes Passenger (Adigüzel, 2016)

2.4.2 Ecological Factors

From the ecology’s point of view, it is interesting who will be responsible in case of an accident. Today the driver is liable for his doings. In automated systems the responsibility will relocate on the automobile manufacturer. The producer has the authority that the vehicle doesn’t cause accidents.

Another factor is that automated systems are a chance for the automobile makers. The earlier they are able to develop automated systems the more they can sell of their products because today very less automobile makers can produce such self-driving vehicles.

2.4.3 Economic Factors

One of the biggest economic factors is efficiency. The automated system can drive in a way it never needs a brake. No energy is unnecessary used because the system knows when which speed is allowed to drive and changes the throttle control on time. Due to this control 15 per cent of fuel can be saved. Automated systems also can drive smoother because people brake more rapidly than it’s necessary. Another aspect is that electric cars can load themselves. For example, during the driver goes shopping the car can drive to the next charging station and charges its battery. No time gets lost for the driver. (Wang, 2015)

2.4.4 Challenges

A direct impact of automated vehicles is the loss of driving-related jobs in transport industries. There also would be needed less work for car services or crash repair shops and in public transport industries.

Privacy could get lost and the risk of hacking could rise as a consequence of the connectivity of the cars. The risk of terrorist attacks could also increase if terrorists control a car loaded with explosives.

The technical realisation has requirements which have to get hold. The most important point at automated driving is the safeguarding against failure. The system is not allowed to fail and stop working. In worst case the driver has to get informed to take over the control. Furthermore, the transmission performance has to work correctly and send the information in real-time. The sensors and actors have to work with the needed accuracy to guarantee the functional compliance.

2.5 State of the Art of Autonomous Driving

Automobile manufacturer like Audi, BMW, Mercedes and Tesla are the leaders in producing automated vehicles.

Till 2013 the only system which was used, was the speed control.

Since that year systems like the hold lane assistant, change lane assistant and park assistant (partly automated) have been developed and used in cars. Hold lane assistance is used for holding the lane on straight streets just as on curved streets on the motorway. Change Lane assistance is needed for the case that a car in front of the automated car drives slower. It changes automated the lane and passes the slower car. After this manoeuvre it changes the lane again to the right lane. (Holzer, 2016)

2014 the safety brake assistant with pedestrian detection was evolved. This system brakes automated if the car would collide with a person or a cyclist. It also stops the car in city traffic if the forward car stops and the driver doesn’t react in time. (Auto, Motor und Sport, 2014)

Since 2015 systems like construction area assistance and the evasive assistant have been developed. Construction area assistance can steer the car through tight lanes and break the car if another car leaves its lane. Evasive assistant is used to avoid a collision to a vehicle, which suddenly enters the street as you can see on Figure 2.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2: Collision Avoidance to a Vehicle (Auto, Motor und Sport, 2014)

2016 a jam assistant which breaks, steers and accelerates automated is the highest automation which is allowed. This system orientates on road markings and also with radar to run ahead of other cars. The driver has to put his hands on-again-off-again the steering wheel otherwise the vehicle will make an audio warning and stop the car with emergency flasher. (Auto, Motor und Sport, 2014)

3 Technical Realisation

Until today the human must watch the environment of the car. The driver has to look ahead and around the vehicle. In an automated system this detection has to work by sensor systems. Now the car itself must monitor as exact as possible what happens around the car. The systems have to realize what the lane of the street looks like, whether there are any traffic signs or traffic lights. (Rauch, Aeberhard, Ardelt, & Kämpchen, 2012)

3.1 Sensor Systems

Much different systems could be used for the detection of possible obstacles or other vehicles. For every use case another system will be the best and it is important to find out which one it is. The most marketable systems will be faced here. (Rauch, Aeberhard, Ardelt, & Kämpchen, 2012)

Abbildung in dieser Leseprobe nicht enthalten

Figure 3: Sensor Systems in an Autonomous Car (Rügheimer, 2016)

3.1.1 Ultrasonic Sound

These sensors are one of the eldest systems (20 years) to check if an obstacle is detected. It’s needed to park the car without crashing other cars or other barriers – with or without a driver on board. Another use case is the detection if another vehicle is in the blind angle of the car. Nowadays these sensors can be used for high-resolution three-dimensional environment detection. Objects at a range of 30 to 300 centimetres can be identified at a precision of 5 to 10 millimetres. (Rügheimer, 2016)

Abbildung in dieser Leseprobe nicht enthalten

Figure 4: Ultrasonic Sound Sensors for detecting Obstacles and Vehicles (smpower)

Functionality

Ultrasonic sound sensors send sound waves higher than the human ear can notice. This waves get mostly reflected, partly also absorbed. The reflected waves get detected by another sensor group. With the phasing and the amplitude, the time between sending and receiving, the signal can be determined. As a consequence, the distance to the object can be calculated. (Rügheimer, 2016)

Table 1 Advantages and Disadvantages of Ultrasonic Sound (Sensata, 2007)

Abbildung in dieser Leseprobe nicht enthalten

3.1.2 Cameras

Cameras are the most important type of sensor systems in an automated vehicle. If systems have to “watch” a large distance cameras are the first choice. Normally a camera is installed in the rear-view mirror in driving direction to see occurrences in the front. The detection of the traffic lane is achieved by cameras to identify traffic signs and hold the lane automatically. Moreover, obstructions can be recognised by this system.

If a camera system has to measure the distance to the front vehicle it works better if there are two cameras (stereo cameras). By means of an algorithm it can get calculated exactly. (Sensata, 2007)

Abbildung in dieser Leseprobe nicht enthalten

Figure 5: Brake Assistant with a Stereo Camera (Wikipedia, 2016)

Functionality

The stereo camera has two CMOS-cameras integrated in a distance of only 12 centimetres. On-board is also the control unit for these two cameras. This module is assembled in the rear mirror in driving direction of the car and doesn’t compromise the view to the street. The cameras are able to capture a field of vision with an angle of 50 degrees and a 3D-operating distance of about 50 metres. By reason of two cameras it’s possible to measure the distance to the vehicle driving in front of the car. This process is filmed with a resolution of 1280 x 960 pixels. It has the safety standard ASIL-B referred to ISO-Norm 26262. (Kuss, 2015)

Abbildung in dieser Leseprobe nicht enthalten

Figure 6 Stereo Camera made by Bosch (Denton, 2015)

Table 2 Advantages and Disadvantages of Cameras (Sensata, 2007)

Abbildung in dieser Leseprobe nicht enthalten

3.1.3 Radar (Radio Detection and Ranging)

Radar means Radio Detection and Ranging and is used for the autonomous emergency braking, the adaptive cruise control and the collision mitigation. By way of its operating distance of 150 metres the car is able to capture information very early. It measures the distance to the other vehicles and if there are any obstacles. Furthermore, it works in the case of bad weather when it’s rainy or foggy. Due to radar the car can have a look through on the contrary to humans. (Balzer, 2014)

Abbildung in dieser Leseprobe nicht enthalten

Figure 7 Front Radar Sensors on a BMW 530 (Group, 2007)

Functionality

Radar works with electromagnetic waves and the evaluation of its echo. A sender transmits the radar waves to the environment. If there is another vehicle or an obstacle, the waves gets reflected back to the sender. Near the sender there is a receptor, which receives the signal. With this echo and the help of the Doppler effect, the distance and the speed of the other vehicles can get calculated. (Wolff, no date)

Table 3 Advantages and Disadvantages of Radar (Sensata, 2007)

Abbildung in dieser Leseprobe nicht enthalten

3.1.4 Infrared

The Infrared-system works in cooperation with the camera. Infrared headlights are obstructed in the car. Although these light sources are invisible for humans they are not for the camera. It shines for an operating distance up to 150 metres and doesn’t blind somebody. The camera can detect pedestrians or animals on the street better than other systems. If any obstacles are detected the car informs the driver immediately on the head-up display or a monitor in the central console as you can see on Figure 8. (Balzer, 2014)

[...]