Prototypes selection by multi‐objective optimal design: application to a reconfigurable robot in sandy terrain

Purpose – Wheel‐terrain interaction has hardly been taken into consideration in the process of conventional mobile robot design, but its importance has been reflected increasingly towards these categories of mobile robots in rough sandy terrain or obstacle‐dense ground, as the first performance index in this situation is the trafficability of robot whose propulsion is uniquely generated by wheel‐terrain interaction. Consequently, it is valuable to find an optimized design method when the terrain and robot itself are regarded simultaneously. The purpose of this paper is to present a novel and reasonable design approach to mobile robot in sandy terrain. Design/methodology/approach – Leading to some conflicted performance indices of robot, terramechanics describes the non‐linear characteristics in wheel‐terrain interaction mathematically, therefore, trade‐offs must be implemented to get a proper solution by multi‐objective optimization (MOO). In this paper, a five‐wheeled drive and five‐wheeled steering (5WD5WS) reconfigurable mobile robot is taken as demonstration with taxonomy of total‐symmetrical, partial‐symmetrical and asymmetrical prototypes. After function modeling, the MOO is carried out via iSIGHT‐FD using NCGA (Neighborhood Cultivation Genetic Algorithm) to minimize the mass, wheel resistance and maximize the static stability simultaneously. Findings – After MOO, a compact and light weighted asymmetrical prototype is obtained with better trafficability, and other prototypes can produce diversified configurations to meet specific requirements. Significantly reduced masses (about 17 kg) enhance the grade‐ability when robot is in rough terrain. Performed real‐world experiments have also verified these prototypes. Originality/value – The paper presents a new design approach for a mobile robot which focuses on both robot and terrain simultaneously with respect to conflicted factors. To unveil the insight relation of these factors, MOO is an effective tool to get a trade‐offs prototype.