Cascaded Sliding Mode Observer-Based Active Tilting Control for Unmanned Narrow Reverse Tricycle in Uneven Terrain

Although narrow vehicles, such as bicycles and motorcycles, exhibit superior traversability and maneuverability on narrow and uneven terrain compared to wider vehicles (e.g., cars, vans, and trucks), their higher center-of-mass-to-track ratios increase rollover susceptibility. To address the demand for unmanned locomotion in such environments, this work proposes an unmanned narrow reverse-tricycle vehicle (UNRTV) with only three actuators: steering, active tilting, and rear-wheel drive. A parallelogram multilink front-body mechanism is designed to address the stability-traversability tradeoff inherent to vehicle width. A dual-layer estimation framework, the sliding mode observer-cascade observer (SMO-CO), is developed for cooperative estimation of system states and external disturbances. A dynamic algorithm for desired roll angle generation, based on state observation, is proposed alongside a state-tracking control strategy incorporating disturbance feedforward compensation. Experimental validation across three scenarios 1) mobility on even terrain, 2) turning on a counter-gradient slope, and 3) locomotion on uneven terrain confirms the engineering feasibility of both the UNRTV platform and the SMO-CO control algorithm. This research presents a novel solution for unmanned logistics in constrained terrains; the developed technical approach further supports the advancement of extraterrestrial exploration vehicles (e.g., lunar/Martian rovers), thus enabling progress in planetary mission technologies.