Research

Project Number:
04-03

Research Project:
Evolution of Collective Sensory Systems for Intelligent Vehicles

P.I. Name & Address:
Karl-H. Grote
California State University, Long Beach
Department of Mechanical and Aerospace Engineering
Long Beach, CA90840
Tel: (562) 985-1525
Fax: (562)985-4408
Email: kgrote@csulb.edu

Project Objective:
Newer passenger cars are equipped with sensors usually in the bumpers of the cars, helping the driver to park the car in close spaces and avoid contact to structures. Some manufacturers also propose and even offer sensors, which allow the driver to adjust the cruise control to a certain distance to the moving car in front.The objective of the proposed research goes further: In order to avoid cross - lane accidents, e.g. on California Freeways, sensors must have intelligent controllers and software to maximize success of driver actions. The sensors have to recognize and evaluate the potential danger in the dynamic traffic environment and give the driver appropriate warnings or even override the driver, in order to prevent an accident.

The design and development in all areas is based on the creativity of the engineer. Since 1960 evolutionary computation methods such as genetic algorithms have been researched and applied to find near-optimal solutions, which can then be used to generate the desired outcome. The previous research applied the evolutionary method to optimize number, position and type of sensors on intelligent vehicles simulated in a traffic environment [1-4]. The evolutionary approach was verified within the framework of a simplified case study, where the cost and coverage of the sensory system were the major competing factors for suggesting its configuration design. And it was also applied to another case study problem, where robotic controllers based on artificial neural networks (ANN) were evolved by the same evolutionary methodology.

The proposed research will continue to focus on research of the evolutionary methods to generate intelligent sensory systems that could analyze the traffic scenario and assist the driver to avoid collisions. Computer simulations will be major parts of the research. The outcome of the research will suggest appropriate hardware and software sensory systems for cars and trucks on multi-lane streets and freeways to avoid contacts with other vehicles or objects. The sensors available for the automotive industry also have to be researched and classified in terms of their characteristics and costs, in order to simulate a realistic design selection process of the automotive industry.

The research will be conducted at CSULB and Caltech. Caltech offers access to simulator and computational facilities. CSULB will conduct the research for the design and methodology for sensor abilities and costs, as well as hardware - compatibility to selected vehicles.

Task Descriptions:
1. Continuous GA/GP research (12 months)
2. Design/selection of intelligent sensors (5 months)
3. Redesign and refinement of sensor technology (6 months)
4. Report writing and final discussion of research results (1 month)

Milestones, Dates:
January 5, 2004 - January 4, 2005

Total Budget:
$35,049

Student Involvement:
One Student @ 20% time 12 months
One Student @ 60% time 12 months

Relationship to Other Research Projects:
New topic area, part of safety and security focus area

Technology Transfer Activities:
Project report posted on the website

Potential Benefits of the Project:
Reduction of highway crashes

TRB Keywords:
4b.12 Collision avoidance; vehicle sensors; highway safety

Primary Subject:
Vehicle safety

Goals:
safety

Enabling Research:
Computer information, commutation

Modal Orientation:
Highway