Adaptive Gait-Based Control for Assistive Robots Supporting Elderly on Inclined Surfaces

This paper presents an adaptive human -robot interaction framework designed to assist elderly individuals in navigating both flat and inclined surfaces using a mobile robotic platform. The proposed method integrates real-time gait estimation and slope detection to dynamically modulate the velocity of the robot, thereby ensuring stability and safety during locomotion. A hybrid sensor architecture, comprising ultrasonic distance sensors and force-sensitive handles, enables accurate estimation of human gait parameters, such as cadence, step length, and center of pressure variations. The inclination angles are computed through geometric analysis of the sensor arrays mounted on the robot, with real-time calibration techniques implemented via FPGA-based hardware acceleration for low-latency response. A torque and velocity control algorithm, informed by biomechanical changes in gait and user-applied interaction forces, adjusts the robot's behavior to match the user's walking intent on controlled planar inclines. Experimental validation was conducted with 30 participants on slopes of up to 40°, demonstrating effective gait synchronization, reduced velocity mismatch, and improved safety through stability-aware motion control, including comparative evaluation against baseline controllers (Section IV-D). The current validation is limited to controlled, rigid ramp surfaces (0°-40°). The manuscript scopes claims accordingly and explicitly discusses ultrasonic sensing limitations on soft/irregular surfaces and non-planar terrain.