Skip to main content

MIEDT: Modeling and Implementation of Expert Driving Techniques towards the Development of New Active Safety Systems for Passenger Vehicles

Final Report Summary - MIEDT (MIEDT: Modeling and Implementation of Expert Driving Techniques towards the Development of New Active Safety Systems for Passenger Vehicles)

In this project we studied driving techniques used by expert race drivers to control their vehicles under extreme operating conditions. We envision that expert driver knowledge can be modelled mathematically and implemented towards the development of a new generation of active safety systems for passenger vehicles. These new systems will employ expert driving skills to assist the driver, or autonomously control the vehicle close to the limit of its handling performance. In this project we focused on driving techniques used by rally-race drivers, who clearly operate beyond the limits enforced by current active safety systems.

With the above vision in mind we set out to fulfil the following objectives:
- Collect expert driver knowledge in the form of empirical guidelines on the execution of expert driving techniques used in rally racing, as well as in the form of driver control commands and vehicle response data during the execution of such expert driving techniques by race drivers.
- Based on this newly acquired knowledge on rally driving techniques, design a control scheme, which uses driver control inputs (steering, throttle and brake commands), to reliably replicate the expert driving techniques recorded, and implement the control architecture in realistic simulation environments.

Collaboration with a rally driving school in the UK was established early on in the project. The rally school provided valuable empirical guidelines on rally driving, in addition to vehicles for testing, technical assistance during the instrumentation of the vehicles, access to their facilities (workshop and testing grounds), as well as expert drivers to perform the manoeuvres and driving techniques during the data collection experiments. Testing facilities and equipment were acquired and developed according to the data collection requirements of the project. A comprehensive sensor and data logging suite was compiled, consisting of a dual antenna GPS receiver, an inertial measurement unit, a CAN-bus interface to collect signals form factory installed sensors on the vehicle (e.g. wheel speeds, engine speed and throttle position), as well as externally fitted steering angle and brake pressure sensors. A base-station providing differential GPS corrections and increasing the position measurement accuracy was finally added to the testing equipment, allowing for studies of the vehicle trajectory with respect to the road limits geometry. A dedicated test vehicle was acquired and fitted with standard safety features (roll-cage, race seats and harnesses) and provided a permanent platform for calibrating the testing equipment and for data collection during the execution of driving techniques associated with Front-Wheel-Drive (FWD) vehicles. Driving techniques associated with Rear-Wheel-Drive (RWD) vehicles were also recorded using a test vehicle from the rally school. Testing of a Formula Student race car in a closed tarmac circuit was also performed for comparison of driving techniques associated with different types of racing.