A Comparison of a Digital and Floating-Point Virtual Programmable Logic Device and an Artifical Neural Network Evolved for Robotic Navigation

Two novel architectures, the digital and floatingpoint virtual programmable logic device (VPLD) are compared with an artificial neural network, evolved for the robotic navigational tasks of obstacle avoidance and light following for a two wheeled robot. The VPLD is based on a programmable logic device but is coded in software rather than in hardware. This allows the VPLD to be implemented on CPUs allowing it to be run on a wide range of platforms including PCs, mobile phones and ARM processors. The function of the VPLD is governed by a configuration bitstream which can be evolved by evolutionary computation. It was found that the digital and floating-point VPLDs performed well against the ANN in the navigational tasks, making the VPLD a viable alternative to an ANN for robotic navigation.