Recent Advances in Autonomous Vehicle Technology-Perception and Path Planning

In autonomous vehicles (AVs), perception and planning are two fundamental components that work together to ensure safe and effective navigation. The perception system interprets the vehicle’s surroundings by processing data from a variety of sensors, including cameras, radar, and LiDAR. This sensor data is used to detect and classify objects—such as other vehicles, pedestrians, cyclists, traffic signs, and road boundaries—and to assess environmental factors like weather and road conditions. Ongoing advancements in machine learning and computer vision play a vital role in enhancing the accuracy and reliability of AV perception. Planning, on the other hand, involves determining the vehicle’s course of action to safely and efficiently reach its destination. This encompasses route planning, behavioral planning, and motion planning. Seamless integration between perception and planning is essential: perception provides real-time, accurate environmental data, while planning systems adapt dynamically based on new inputs from perception. Achieving this integration requires a carefully designed pipeline that ensures rapid, reliable data flow, often utilizing software frameworks that prioritize both low latency and high dependability. This book explores the latest technological advancements in both perception and planning, with a focus on their integration. It also delves into cutting-edge developments in autonomous racing, a domain that pushes the limits of AV technology. It is intended for a broad audience, including academic researchers, graduate students, automotive engineers at OEMs and suppliers, software and ICT professionals, as well as managers and decision-makers in the field.

The Content

• Scenario-based quantification of the impact of automated vehicles on traffic
• Adversial examples in environment perception for automated driving
• Stable resampling strategies for radar-based dynamic occupancy grids
• Local point cloud features for LiDAR self-supervised representation learning
• On torque-vectoring control for the obstacle avoidance scenario: a reinforcement learning case study
• Performance evaluation of path tracking controllers for scaled robotic cars
• A nonlinear dead-time compensation method for path tracking control

The Editor

Daniel Watzenig was born in Austria and holds a doctorate in electrical engineering. He received the venia docendi (habilitation) in electrical measurement science and signal processing from Graz University of Technology, Austria. He currently serves as Chief Technology Officer (CTO) and Head of the Electronics Systems and Software Department at Virtual Vehicle Research, Graz. In addition, he was appointed Full Professor of Multi-Sensor Perception of Autonomous Systems at the Institute of Visual Computing, Faculty of Computer Science and Biomedical Engineering, Graz University of Technology. His research interests include sensing and control of autonomous vehicles, sensor fusion, reinforcement learning, and decision-making under uncertainty. Professor Watzenig is the author or co-author of more than 200 peer-reviewed publications, book chapters, patents, and articles. He is the Editor-in-Chief of the SAE International Journal on Connected and Automated Vehicles (SAE JCAV). He has been a guest lecturer at Stanford University, USA, since 2019, teaching multi-sensor perception for autonomous systems. In 2023, he also began teaching robotics classes as a guest lecturer at Tongji University in Shanghai, China. He is the founder of the Autonomous Racing Graz Team and, since 2024, has served as Vice Chair and Executive Committee Member of the IEEE Austria Section. He is also an IEEE Distinguished Lecturer in the field of autonomous vehicles, a Board Member of the INSIDE Industry Association (European Initiative on Intelligent Digital Systems), and a Member of the Academic Advisory Council of PAVE (Partners for Automated Vehicle Education). Since 2019, he has acted as a consultant and appointed expert in military robotics for the Defense Technology Agency of the Austrian Armed Forces.

The Book Series

Sustainable Mobility and Energy highlights the crucial role of sustainable mobility in reducing CO2 emissions. This peer-reviewed series features insights from leading experts on electric and fuel cell vehicles, which offer eco-friendly alternatives to conventional engines, especially when powered by renewable energy sources. Utilizing clean energy significantly reduces air pollution and dependence on fossil fuels. Additionally, the integration of modern communication technology enhances traffic control, reduces congestion, and increases road safety. Join us in exploring innovative solutions for a greener and more efficient future.