Instance-based Learning of Vehicular Performance Models

Abstract

Motion planning with predictable timing and velocity will enable a number of interesting applications such as autonomous intersection management. These planning algorithms depends on an accurate model of the performance of the vehicular controllers, which can be highly non-linear. Au et al. proposed a motion planning algorithm that is based on a performance model of the vehicle controllers. However, they assumed that the performance models are given for every road, and did not discuss how to build these models. In this paper, we propose an instance-based learning method to learn a performance model automatically, and compare it with the artificial neural networks. We argue that instance-based learning is suitable for this task because performance models for different roads are quite similar. An exploration strategy based on the principle of least effort is given to speed up the learning process. Our experiments showed that the instancebased learning method with distance-based exploration strategy has a faster rate than the artificial neural network methods.