On-Road Intelligent Vehicles

Motion Planning for Intelligent Transportation Systems

Rahul Kala

• Editorial: Butterworth-Heinemann
• Año: 2016
• ISBN:
  9780128037560
• Categoría: Technik & Maschinenbau
• Categoría: Automobil- & Transportbranche

On-Road Intelligent Vehicles: Motion Planning for Intelligent Transportation Systems deals with the technology of autonomous vehicles, with a special focus on the navigation and planning aspects, presenting the information in three parts. Part One deals with the use of different sensors to perceive the environment, thereafter mapping the multi-domain senses to make a map of the operational scenario, including topics such as proximity sensors which give distances to obstacles, vision cameras, and computer vision techniques that may be used to pre-process the image, extract relevant features, and use classification techniques like neural networks and support vector machines for the identification of roads, lanes, vehicles, obstacles, traffic lights, signs, and pedestrians.

With a detailed insight into the technology behind the vehicle, Part Two of the book focuses on the problem of motion planning. Numerous planning techniques are discussed and adapted to work for multi-vehicle traffic scenarios, including the use of sampling based approaches comprised of Genetic Algorithm and Rapidly-exploring Random Trees and Graph search based approaches, including a hierarchical decomposition of the algorithm and heuristic selection of nodes for limited exploration, Reactive Planning based approaches, including Fuzzy based planning, Potential Field based planning, and Elastic Strip and logic based planning.

Part Three of the book covers the macroscopic concepts related to Intelligent Transportation Systems with a discussion of various topics and concepts related to transportation systems, including a description of traffic flow, the basic theory behind transportation systems, and generation of shock waves.

• Provides an overall coverage of autonomous vehicles and Intelligent Transportation Systems
• Presents a detailed overview, followed by the challenging problems of navigation and planning
• Teaches how to compare, contrast, and differentiate navigation algorithms

1

Introduction

Abstract

The vehicles are increasingly becoming intelligent and can do many of the tasks that otherwise require a human driver, including the ultimate level of completely driving themselves. Small considerations may hinder an immediate widespread adoption, but the rise in the technology leading to a widespread adoption is very eminent. These vehicles enable efficient driving, while freeing the driver to do more pleasing and productive activities instead of driving. The vehicles can be seen as mobile robots which operate on the road, while using advanced artificial intelligence algorithms for navigation and decision-making. Future makes it possible to create ad hoc networks of vehicles, including the roadside units and transportation management centres for more informed decision-making. This paves the way for a fascinating future in which robots of advanced capabilities will occupy the transportation network, the transportation network itself being capable of efficiently and safely handling traffic of high speed and diverse capability.

Keywords

Autonomous vehicles; Intelligent transportation systems; Intelligent vehicles; Inter-vehicle communication; Mobile robotics; Planning

1.1. Introduction

Standing at the cliff of a remarkable technology, and to see it change the very fundamental picture of the most important aspect of life is similar to experiencing the majestic history being created in front of you. We have already travelled the era of discovery of automobiles, public transport, railways, ships and airways, and witnessed the impact these technologies have had on the life of one and all. We have already witnessed the automobiles scale up to the most economical scales, causing less pollution, driving at the highest speeds and levels of comforts and at the same time displaying a very high level of luxury. As a result, it is common to see people passionate about the vehicles they possess and love driving or travelling with these vehicles.

The era of the human-driven vehicles being eliminated and being replaced by the self-driving cars is almost here, with some connotations attached. These vehicles have a lot to offer including driver ease, no need of hiring a driver, ability to sleep and enjoy while the machine drives, ability to move efficiently, not worrying about the parking etc. Therefore, the technology is irresistible. This also gives the ability to frame the next generation and extremely efficient transportation networks, which will enable the quickest and safest transportation of a large number of people, operating using futuristic technologies, the most sophisticated Artificial Intelligence tools and techniques and using hardware and software solutions like never before.

It is intriguing to open the wrapper of this interesting technology to get an understanding of the complex concepts which make it happen, while presented in manner simple to understand and easy to master. This book is devoted to the same. Fundamentally, the autonomous vehicles can be seen as mobile robots, which can look around, understand the traffic scenario, make intelligent decisions and act upon them, exactly similar to the way in which humans drive the vehicles. So one needs the basic hands, legs and body to be found in the autonomous vehicles; mechanisms to understand what the eyes see and mechanisms to instruct the legs to move as desired; and some memory to remember the happenings. Once we get these building blocks, the machines can be taught how to drive, like a driving instructor does for the humans, which is the role that we will assume while designing all the planning algorithms. The ability to instruct is largely facilitated by the available tools and techniques of Artificial Intelligence, which though need to be tweaked to suit the specific requirements.

Once the vehicles are taught how to drive, their role can then be extended to traffic inspectors to decide how the traffic should operate and to regulate the traffic accordingly, thereafter designing algorithms which automatically do so. Ultimately, the role is extended to the overall transportation manager who knows the states, sources, destinations, cause of travel etc. for all vehicles and has the potential to regulate traffic so that the overall goodness is maximized. The purpose of the book is to design efficient algorithms which do all of this. It is important to realize the potential of having all vehicles talk to each other, additionally talk to the intelligent agents located at the roadside and the central transportation management authority.

The current demonstrations by academia, research labs and other companies showcase the promising future of the technology. The visitors to these demonstrations normally find it hard to believe their eyes that a vehicle is indeed driving on its own, and that the demonstration is not a scary movie or a nightmare. The people who actually get a chance to sit inside are even more excited, although a bit scared, not immediately trusting the technology. These demonstrations ultimately suggest a very soon and a very effective widespread adoption of these vehicles, with some caution and concern. With such a background, it is natural to extrapolate the developments in the related domains of robotics, vision, intelligent vehicles and intelligent transportation systems. The future will see extraordinary vehicles, engineered by unique designs suited to the specific task that they will be expected to perform, operating in the next-generation transportation systems, in which every imagination is possible. It further goes to eliminate the gap between the on-road and off-road vehicles, mobile robots and autonomous vehicles and, more surprisingly, the land, air, water and underwater robotics. In the future one does expect robots to have extraordinary capabilities, similar to the ones being widely projected in different science fiction movies. The book presents the in-depth technology behind the intelligent vehicles and intelligent transportation systems, in pursuit of designing the futuristic transportation networks.

1.2. Why Autonomous Vehicles?

Before delving deeply into the technology, it is always better to assess the end returns. Autonomous vehicles have a lot to offer in the future and clearly fascinate young researchers and technocrats, which is motivating enough to pursue research and development in the area. It is easy to portrait oneself as an architect of the future transportation system, to sculpt each and every aspect of the vehicle and the overall system. The technology behind autonomous vehicles and intelligent transportation systems provides the enabling tools to showcase such a future. Although the pros may be highly benefiting and rewarding, one must never underestimate the cons and the costs incurred or likely to be incurred in the long run. This section puts the pros and cons of the technology into perspective, to clarify the context before the book takes the readers deeper into the insight about the technology.

1.2.1. Advantages

Autonomous vehicles are largely motivated by passenger safety. Every year numerous people die or are seriously injured in road accidents. The accidents are largely caused by human errors in perception and decision-making, or by lack of attention and sometimes by late reactions. The autonomous vehicles can use the active percept of a number of redundant sensors to timely make correct driving decisions and hence eliminate accidents. Autonomous vehicles further result in more efficient travel. This means one has to spend much less on insurance, while the insurances will naturally become less expensive. The vehicles benefit from efficient onboard algorithms, communication with the vehicles around, the transportation management centres and information from different sources. This facilitates making better informed and optimal decisions. Further, these vehicles can park themselves close to each other and operate in high-speed platoons, which significantly increase the operational efficiency. The implications of a single vehicle may seem limited, but a network of such intelligent vehicles has the potential to make the overall transportation system operate efficiently. As an example, the vehicles can collectively operate to avoid congestions.

Driving is not always a pleasing and joyful experience and many people have to force themselves to drive. The autonomous vehicle relieves the humans from this mundane task, and the person can instead let the vehicle drive itself and do tasks that are more important, take rest or simply entertain one. This is especially important from the point of view of social justice, as these vehicles will enable the disabled and elderly citizens to commute who currently cannot on their own. It is also possible for vehicles to deliver goods and services without any human onboard. This facilitates the use of autonomous vehicles for delivery of courier, goods, automated taxis, continuous long-distance travel, as a means of public transportation, as guide vehicles, inspection vehicles etc. The vehicles can also be used to negotiate difficult-to-drive roads, which is not very easy for the humans to negotiate. Many times the human drivers need to be hired, which increases the cost of transportation, which means that both time and operating costs are reduced by the use of autonomous vehicles, resulting in less expensive and better services. Not only will the manpower be reduced, the number of vehicles will be reduced as well. The vehicle will be able to leave the passenger, park it and come to pick up the passenger on being asked remotely. This will result in fewer parking spaces needed around the core areas of markets and offices.

1.2.2. Concerns

The picture is not entirely beautiful; there are some black spots as well, although the limitations and concerns are much less compared to the advantages. The greatest concern in the use of autonomous vehicles is the legal and legislative aspects. These vehicles are already giving a hard time to the legislators to frame policies about their operation. An accident by a human-driven car has well-defined laws about the defaulter and the punishment, but in case of an accident by an autonomous vehicle, naturally the vehicle itself cannot be punished for poor driving. The issue raises serious ethical concerns, because a robot is not allowed to harm or kill a human, which may be possible in case the algorithms of the autonomous vehicle falter. It is still unclear how the vehicle should react when an accident is imminent, and a choice has to be made whether to save the passenger, an infant or a grown-up pedestrian. Humans are able to make such decisions, whereas vehicles may not have the same ethical judgement ability.

Costs are a matter of great concern. Autonomous driving capability adds costs to a traditional vehicle in terms of both hardware and software. Because a lot of money is invested in research and development, the initial software costs are expected to be very high. Lack of acceptance by the initial market, added to high costs, may diminish the acceptability of these vehicles. The hardware costs are as well currently very high, which need to be scaled down to be accepted. These vehicles will create sources of automation, and, as a result, numerous drivers of trucks, buses, taxis etc. will become jobless. This will create unemployment unless an alternative source of employment is generated. The technology already faces the criticisms of trade unions and workers.

There are some implementation issues as well. These vehicles always obey laws, whereas human drivers often break laws by some magnitude. One may not always make optimal separations with the vehicles in front, make close cut-ins and overtakes, get inside a roundabout even though it may not be very safe etc. This causes the autonomous vehicles to under-perform in comparison to the human drivers. Although teaching the vehicles to break traffic laws is not suggested, operating in such scenarios questions the acceptability of these vehicles. They may wait for a prolonged time at roundabouts or cause trouble to the drivers around who expect the vehicle to drive a little more aggressively. Further, the autonomous vehicle owners will constantly want to reduce the number of human-driven vehicles due to the added benefits; the people will continue to demand their right to drive on their own. This will cause debates in the future, especially considering that some people have a passion for driving and are not willing to let the vehicle drive itself. It is and will always be uncertain if these vehicles have been trained against all kinds of traffic and all kinds of traffic situations. This creates a constant fear of accident due to unseen scenarios not considered in design. Because everything is networked, privacy is another concern; so is the case of the hacking of networks.

1.3. A Mobile Robot on the Road

Intelligent vehicles operating autonomously on the road may be easily seen as mobile robots which navigate from one place to another. The fundamental technology behind autonomous vehicles is the same as that of mobile robotics (Holland, 2004; Siegwart et al., 2011). Both of the techniques have the problem of navigating a robot from one place to another, while avoiding collisions with other robots and entities which may otherwise be treated as static or dynamic obstacles. The output is a trajectory to be followed, which must be the shortest, fastest or the safest to be called optimal. Autonomous vehicles are a special example of the general class of mobile robots. Like mobile robots, these vehicles perceive the world, make a map representation of the percept which is constantly updated, plan the motion in multiple layers of hierarchy and finally move as per the plan. Hence, the techniques discussed in the book will be, first, from the perspective of general mobile robotics and then noting the differences that arise when the same techniques are applied for the specific problem of autonomous-vehicle driving. Mobile robots largely operate on widely bound or unbound home and office environments, whereas vehicles have a well-defined road structure and need to obey the traffic laws and best driving practices. This can be both a boon and a curse to the problem of vehicle navigation, in contrast to the general problem of the navigation of mobile robots. Throughout the book, this aspect of the difference will be highlighted.

1.4. Artificial Intelligence and Planning

Many breakthroughs in the domain of robotics have happened due to the development of Artificial Intelligence (AI) tools and techniques (Konar, 1999; Russell and Norvig, ...