Functional and modular organization of planning subsystems of mobile robot behaviour with partial uncertainty for the two-dimensional space

Recently we can see significant increase in consumer demand for autonomous mobile robots that are used for a wide range of functional tasks in a not completely determined environment. The actual task that we consider is solution problem of development of an effective situational planning subsystem for the behaviour of autonomous mobile robots. The purpose of writing this article is a brief coverage of basic principles (proposed by the authors) of functional and modular organization of situational planning subsystem, that is included in system of position trajectory control for mobile robots. The situational planning subsystem is designed to determine the values of the parameters of the safest movement in environment with moving and stationary obstacles to the goal. Designed situational planning subsystem structural is differed from others by using of neural network basis in the construction of a situational planning subsystem functional kernel. The presence of built-in detection means and short-term auto-tracking trajectories of movable obstacles with considering a priori to certain limitations, the calculation of the spatial and velocity zones where collisions with the obstacles are the most possible are differed our planning subsystem too. Also we determinate parameters of mobile robot manoeuvre on the plane, which in this case should be perform. In the basis of functional and modular organization for situational planning subsystem was put situational planning method that was developed by the authors. We designated it CDVH-NN (distance vector histogram — neural network complex method). Integration of algorithmic means of short-term forecasting of the external environment with the formal-logical basis of direct distribution neural networks are different CDVH-NN method from others. The productivity and the efficiency of the functional and modular organization of situational planning subsystem are confirmed by the results of software simulation in MatLab and executed natural experiments. In the article, we present results of modelling of CDVH-NN kernel in basis on FPGA for family Altera Cyclone IVe. We note that the implementation of situational planning subsystem using SOPC technology increase on several times of situational planning subsystem performance with comparison to its full program implementation. It allows freeing up resources of mobile robot board computer for other important operational task.