Self Driving Vehicles and Risk Analysis
Guidance, Navigation, and Control (GN&C) systems take data from the environment to create a mathematical representation of the system state. The system’s guidance solution is then determined from this navigation solution and is typically a minimization of some objective weight function, J. This objective function is a time, cost, weight, or other physically realizable quantity. We show a method to create a guidance solution based upon system risk for a system’s goal. The methodology presented uses a goal to create a situation. The situation assessment creates a situation model whose risk can be determined. This risk assessment can then be learned by pattern recognition systems to make decision on how to control a vehicle. This is begun by looking at how to transfer knowledge of a lower complexity problem into increasingly more complicated control problems. Using this method, work has been done to develop a small robotic car to drive autonomously between "lane lines" using only visual information.