Analysis and evaluation of a real-time horticultural autonomous vehicle system

This paper deals with the analysis and evaluation of the real-time software of an autonomous vehicle system, developed at the Silsoe Research Institute. In order to achieve the goals of this autonomous vehicle, the system acquires and processes image and odometry information, using a transputer network architecture. These system functions are represented through the Perts task model which is evaluated using a simulator called DRTSS. The analysis phase sets up the model task and the assignment of their attributes; the evaluation phase provides several metrics to obtain the response time of the computation system and the utilisation rates of its resources. The results of this evaluation will help to decide alternative system implementations