Project topics and materials
Get Complete Project Material File(s) Now! »
Evolution of the ITS to the automated vehicle
Economic and social impact of automated driving
This subsection presents some remarkable figures and facts about the social and economic impact of the ADAS and the automated vehicles, based on their current status of penetration on the automotive market, as well as on short-term predictions.
The cumulative safety contribution of available ADAS technologies works out to $16,307 per vehicle over a vehicle’s 20-year life [Mosquet et al., 2015]. If all new-car buyers made an investment of $8,240, which is the price of these features, it would reduce by 30% the number of crashes and by 9,900 the number of fatalities in the United States. Furthermore, the motor crashes cost the USA $910B or 6% of the real gross domestic product each year. Current ADAS features could save $251B annually. Currently, it would represent a 98% of safety return delivered over vehicle’s lifetime, which could become a 439% with the future fully automated cars [Mosquet et al., 2015]. Despite their enormous potential to improve transportation systems, ADAS features have a slow adoption curve, related to the economic cost for consumers. They are unwilling to pay as much for ADAS features as they cost to make and market, as can be seen in Figure 2.3. For example, most consumers appointed that they would pay on average $270 and as much $400 for the Blind Spot Detection system (BSD) when the actual cost is $595 per vehicle.
Figure 2.4 shows the evolution of the ADAS market up to 2016 and the expected evolution up to the horizon of 2020. It would reach up to $60.14 billion by then, registering a Compound Annual Growth Rate (CAGR) of 22.8% during the period 2014-2020. The growing trend for comfort and safety while driving, along with favorable government initiatives has contributed to this growth.
Improving the transportation systems is not only relevant for safety, but also from the economic point of view. Although the highest severity crashes decreased by 16.8% in the last decade in the United States, the number of motor-powered vehicle fatal crashes have increased 7% from 2014 to 2015, with an increment of 4.1% of non-fatal injury crashes and a 3.7% increase in property-damage-only crashes [NHTSA National Center for Statistics and Analysis, 2017b]. On average, 96 people died each day, and one person was killed every 15 minutes in motor vehicle accidents in the United States. Therefore, the estimated economic cost for the material losses that result from all motor vehicle traffic crashes in the USA in 2010 was $242 billion (see Figure 2.5).
The National Highway Traffic Safety Administration (NHTSA) of the U.S Department of Trans-portation (USDoT) pointed out that in 2015 there were more than 32 thousands of fatal motor vehicle traffic crashes, resulting in more than 35 thousands of fatalities [NHTSA National Center for Statistics and Analysis, 2017a]. Indeed, 45% of these accidents and 44% of the fatalities oc-curred in urban areas. Although the rate of urban fatalities has declined by 18% from 2006 up to 2015 (see Figure 2.6), these figures confirm that despite both industry and research have been working on the integration of in-vehicles ADAS it has not been enough to improve transport safety. Additionally, urban areas still suppose a big challenge regarding safety because of the interaction between cars and other vulnerable road users (VRU) such as pedestrian, cyclists or other two-wheeled motorized vehicles. Besides, cities would continue growing, reaching a percentage of 70% of the people living in cities and only 30% in the countryside [Verband der Automobilindustrie e.
Connected vehicles should lead to a reduction in the number of accidents on the roads, achieving a 90% of reduction by deploying more applications in domains such as the road design, traffic management, vehicle design, information and communications technologies, and human systems integration [Barbaresso et al., 2015].
Historical overview of the ITS – research and industrial projects, demon-strations and competitions
Intelligent Transportation Systems (ITS) emerged in the 1970s willing to facilitate the safe, clean, efficient and comfortable mobility of people and goods, saving lives, time and money. ITS are defined as systems that apply telecommunications, electronics and information technologies into road transport to plan, design, operate and maintain such transport systems [Nowacki, 2012]. Figure 2.7 shows the timeline of the main developments in Europe, United States, Australia, and Japan, from their birth up to the acceptance of the ITS term. The first developments in the ITS appeared in the 1960s in the United States with the Advanced Vehicle Control System (AVCS) of the General Motors research group, which provided both automated lateral and lon-gitudinal control. Besides, the MIT launched the METRAN (MEtropolitan TRANsportation) project, whose aim was applying new control techniques to urban transportation. This led to the further conceptualization of the ITS [Dingus et al., 1996], and the foundation of the CACS program (Comprehensive Automobile Control System) in Japan in the 1970s, to test an interactive route guidance system with an in-vehicle display unit in urban areas.
In the 1980s and beginning of the 1990s, the conditions for the developments of ITS were de-termined. The technological development of mass memories made possible a cheaper information process, which encouraged both manufacturers and the European Community to develop concur-rently two projects in Europe: (i) the Eureka PROMETHEUS project (PROgraMme for a Euro-pean Traffic of Highest Efficiency and Unprecedented Safety, 1987-1995) [Eureka, 1987-1995], to improve the competitive strength of Europe by simulating developments in information technology, telecommunications, robotics, and transport technology. (ii) And the DRIVE project (Dedicated Road Infrastructure for Vehicle Safety in Europe, 1988-1991) [European Commission, 1988-1991], a European Commission project which looked forward to a Europe in which drivers would be better informed and in which intelligent vehicles would interact with their surroundings. The European Road Transport Telematics Implementation Coordination Organization (ERTICO) was funded in 1991 for all European and international organizations to work together for the sustainability of transport through the ITS.
At the same time, there were other projects to deploy the ITS worldwide. In Japan, the RACS (Road/Automobile Communication System) project [Takada et al., 1989] in 1984 formed the basis for the current car navigation system. In Australia, the project TRACS (Traffic Responsive Adaptive Control System) appeared as a pioneering project to evolve transportation concerning traffic management systems. In the United States, the Mobility 2000 group was the precursor of the IVHS (Intelligent Vehicle Highway Systems) program, a Federal Advisory Committee for the US Department of Transportation (USDoT).
The ITS America was established as a non-profit organization, and the term ITS was accepted in 1994 [Auer et al., 2016]. Since then, the prior programs started to be implemented and telematics was settled as a major topic of research, intending to develop new ITS applications and defining its standards, as promoted in the IV EU Framework program.
In 1997, the California PATH group, in collaboration with General Motors, presented an eight cars platoon on a highway scenario during the National Automated Highway Systems Consortium (NAHSC) Proof of Technical Feasibility Demonstration held in San Diego [Shladover, 1997]. The platoon operated with an inter-vehicle distance of 6.5 meters, accelerating, decelerating and per-forming coordinate stops, at speeds as high as the full highway speed (around 105 km/h) to prove the feasibility of improving the throughput of the transportation in highways.
Cybercars concept appeared in the 1990s [Parent and de La Fortelle, 2005]. These are fully automated road vehicles designed for passengers or goods transport, operating on-demand and with door-to-door capabilities for short trips at low speeds in urban areas. In 1997, they were operated for the first time in long-term parking at the Amsterdam Schiphol airport. Since then, several European projects (such as Cybercars and Cybercars-2) and different demonstrations have been done in the 2000s and 2010s to introduce cybercars in the cities public transport (such as the ones in La Rochelle or Antibes), considering a fleet of these cars operating together on platoon configurations.
Regarding international competitions, in 2004 the Defense Advanced Research Projects Agency (DARPA) proposed a Grand Challenge in the Mojave Desert region in the USA. It was the first long-distance competition for automated cars in the world, whose goal was to encourage the re-search and development of technologies needed to create the first fully automated ground vehicles capable of completing an off-road course. None of the vehicles participating in the first edition of 2004 finished the route. A second edition of the competition was held in 2005. Then, five vehicles completed the route. In 2007, an urban version of the challenge was held in Victorville, California. There, a mock urban scenario comprising four-way stop intersections, U-turns, and parking areas. There, six of the 11 participating vehicles completed the 90 km course, where vehicles had to respect all traffic regulations, dealing with other traffic and obstacles, and merging into traffic.
Since interoperability is a critical factor for a deeper development of ITS, new challenges sought to boost the development of cooperative vehicles, able to operate together efficiently by exchanging and interpreting data, providing information and services to other systems in real-time, such as the state of the roads, allowing better traffic management. For instance, it would allow an ambulance to arrive faster to the hospital by changing the timing of the traffic lights after notifying the accident. The Grand Cooperative Driving Challenge [Lauer, 2011] took place as an important competition to deepen the cooperative automated driving. It was held on a highway closed to traffic between Helmond and Eindhoven (Holland), in 2011. There, the nine European teams participating had to develop the longitudinal controller for a platoon configuration of heterogeneous vehicles, where they were exchanging their positions, velocities, and accelerations through wireless communication. There, research on new algorithms for sensor fusion, vehicle-to-vehicle communication, and cooperative control was tested [Geiger et al., 2012]. A second edition of the GCDC was held in 2016 as part of the European project i-GAME. On that occasion, three challenging cooperative scenarios were the focus of the competition: automated platoon merge, automated crossing and turn at an intersection, and automated space-making for emergency vehicles in traffic jam [Englund et al., 2016].
In 2010, another remarkable demonstration was carried out by the Vislab group (University of Parma), as part of the VIAC project. It consisted of an international journey with the PROUD automated car from Parma to Shangai. The course combined rural, freeway and urban open roads, where the vehicle was capable of dealing with the public traffic [Broggi et al., 2014].
There exist some other relevant projects that appeared between the 2000s and the 2010s such as HAVEit, SPITS or DESERVE. HAVEit project (Highly Automated Vehicles for Intelligent Trans-port) aimed to contribute to higher traffic safety and efficiency by designing a task repartition scheme between the driver and the co-driving system, a failure tolerant vehicle architecture and developing and validation the next generation of ADAS directed towards a higher level of automa-tion. The SPITS project (Strategic Platform for Intelligent Traffic Systems) was a Dutch project that aimed to improve mobility and safety, focusing on three main areas: traffic management through cooperative driving and mobility, development of an upgradeable in-vehicle platform to deploy the different in-vehicle systems, and a service download and management solution. DE-SERVE (DEvelopment platform for Safe and Efficient dRIVE) European project (2012-2015) aimed to establish a new embedded software and hardware design by using a more efficient development process, overcoming challenges in reducing component costs and development time of future ADAS functions [Kutila et al., 2014].
Finally, from the Strategic Plan for IVHS, six functional areas can be identified in the devel-opment of the ITS [Sussman, 2008]:
• Advanced Traffic Management Systems (ATMS), to predict traffic congestion, offer alter-native routing instructions improving the efficiency of the highway network, being able to control the traffic dynamically and performing an incident detection to reduce the road traffic.
• Advanced Traveler Information Systems (ATIS), to provide data both to road users at their vehicles or their workplaces and to transit users, such as the location of incidents, weather problems, road conditions, optimal routing and lane restrictions.
• Advanced Vehicle Control Systems (AVCS) -now Advanced Vehicle Safety Systems (AVSS)-, to make the travel safer and more efficient, which comprises collision warning features and emergency brake assist. In the longer term, those systems would imply a higher infrastructure information treatment to improve the efficiency of the roads, concept known as Automated Highway System (AHS).
• Commercial Vehicle Operations (CVO), improving the productivity of trucks, vans and taxi fleets.
• Advanced Public Transport Systems (APTS), to enhance the accessibility of information to users of public transport and the scheduling of public transport vehicles.
• Advanced Rural Transportation Systems (ARTS), to face the economic constraints in low-density roads.
Nowadays, these functional areas of the ITS are covered by 31 user services, which surged as an evolution of the National ITS Program Plan in 1995, providing a comprehensive planning reference for ITS, illustrating how the goals of ITS could be addressed through the development of these inter-related user services [Walton et al., 2000].
Advanced Driver Assistance Systems (ADAS)
Advanced Driver Assistance Systems (ADAS) are vehicle-based ITS designed to improve road safety concerning crash avoidance and injury prevention (primary safety), reduction of injury in the event of a crash (crash protection or secondary safety) and post-impact care assistance (to reduce the consequences of injury) [European Commission, 2016a]. They were born as an evolution of the first systems applied for the safety or convenience, such as the cruise control (1958), the seatbelt reminders (the 1970s), anti-lock braking systems (1971) and electronic stability control (1987) [Mosquet et al., 2015]. Nevertheless, as the European Commission pointed out, not all the in-vehicle systems are used for safety purposes but are also intended to improve the comfort or to manage the traffic.
ADAS can also be defined as electronic components installed in modern vehicles that present an intelligent driving experience to the driver. Their main challenge is the green, safe and supportive transportation, in particular to the accident-free mobility scenarios. Thus, the ADAS have three fundamental functions: aid, warn and assist the driver [Mosquet et al., 2015]. These systems began to be commercialized in the 2000s. So far, we can distinguish the following.
Table of contents :
Acknowledgment
1 Introduction
1.1 Motivation
1.2 Objectives
1.3 Manuscript organization
1.4 Contributions
1.5 Publications
2 State of the art
2.1 Evolution of the ITS to the automated vehicle
2.1.1 Economic and social impact of automated driving
2.1.2 Historical overview of the ITS – research and industrial projects, demonstrations and competitions
2.1.3 Advanced Driver Assistance Systems (ADAS)
2.1.4 Levels of driving automation for on-road vehicles
2.2 Functional architecture for automated vehicles
2.2.1 Advanced automated vehicle systems
2.2.2 A reference automated vehicle architecture
2.3 Path Planning Techniques for Automated Vehicles
2.3.1 Graph search based algorithms
2.3.1.a Dijkstra
2.3.1.b A-star (A*) based algorithms
2.3.1.c State lattices
2.3.2 Sampling-based algorithms
2.3.2.a Probabilistic RoadMaps (PRM)
2.3.2.b Artificial potential fields
2.3.2.c Rapid Exploring Random Tree (RRT) and Enhanced RRT (RRT*)
2.3.3 Interpolating curves algorithms
2.3.3.a Straight lines and circular arcs
2.3.3.b Clothoids
2.3.3.c Polynomial curves
2.3.3.d Splines
2.3.3.e Nurbs
2.3.3.f Bézier curves
2.4 Discussion
3 Path planning in static environments
3.1 Problem description
3.1.1 Assumptions and constraints
3.2 Global planning
3.3 Local planning
3.3.1 Pre-planning stage
3.3.1.a Bézier curves based path planning
3.3.1.b Algorithm description
3.3.1.c Optimality criteria
3.3.1.d Validation of the proposed optimality criteria
3.3.1.e Checking the optimality of the evaluated curves
3.3.1.f Human-like driving behavior
3.3.1.g Databases resolution
3.3.1.h Vehicle model
3.3.2 Real-time planning stage
3.4 Conclusions
4 Path planning in dynamic environments
4.1 Overview of dynamic path planning strategies
4.2 Problem formulation
4.3 Dynamic local planning algorithm based on a virtual lane generation and a grid discretization
4.3.1 Grid-based discretization
4.3.2 Virtual lane generation
4.3.3 Re-planning method for dynamic scenarios
4.3.4 Safety avoidance path
4.4 Conclusions
5 Valdiation Tests
5.1 Validation platforms: simulator and vehicles
5.1.1 Simulation tools
5.1.2 Vehicles
5.2 Validation tests for the static path planner
5.2.1 Results on simulation platforms
5.2.2 Results on real platforms
5.2.2.a Cycab experiments
5.2.2.b Citroën C1 experiments
5.3 Validation tests for the dynamic path planner
5.3.1 Results on simulation platforms
5.3.1.a Static obstacles scenario
5.3.1.b Dynamic obstacles scenario
6 Conclusion
6.1 Conclusions and remarks
6.2 Contributions to the state of the art
6.2.1 Static environments
6.2.2 Dynamic environments
6.3 Future work .
