New Field Operational Tests Sampling Strategy Based on Metropolis-Hastings Algorithm

As recently stated by National Highway Traffic Safety Administration (NHTSA), to demonstrate the expected performance of a highly automated vehicles system, test approaches should include a combination of simulation, test track, and on-road testing. The simulation part need to be based on a probabilistic approach. To do so, an appropriate sampling strategy is often used. In this paper, we propose a new sampling strategy based on Markov Chain Monte Carlo (MCMC) methods, using Metropolis-Hastings algorithm to generate samples from probability distributions of Field Operational Tests (FOT); the Safety Pilot Model Deployment (SPMD) in our case. We begin by modeling the probability distribution of each test parameter retrieved from the SPMD database, two estimation methods were applied: Kernel Density Estimation method and EM algorithm. A comparison was made between the two methods to choose the best one. These distribution models are then sampled using our sampling strategy based on the Metropolis-Hastings algorithm.