Design and Preliminary Prototype Evaluation of a Low- Cost IoT-Enabled Stair-Climbing Wheelchair

Abstract This paper reports a preliminary laboratory evaluation of a low-cost stair-climbing wheelchair prototype and identifies the mechanical limits that govern its failures under load on a representative Egyptian residential staircase. The prototype, a smart stair-climbing wheelchair (SCW), combines a rear motorized stair-tread mechanism actuated by two 12 V linear actuators, differential-drive 24 V brushless hub motors, and a NodeMCU ESP8266 control and IoT layer providing Blynk-based remote override, GPS tracking and infrared obstacle detection for level-ground driving. Analytical models of actuator kinematics, actuator force, drive torque, power budget, traction and static stability were developed and compared with measurements. On a five-step laboratory staircase (riser 175 mm, tread 275 mm, pitch 32.5°), the prototype completed 10 of 10 unloaded ascents (100%; exact 95% CI 69–100%) and 7 of 10 ascents with a 70 kg dead-weight load (70%; 95% CI 35–93%). All loaded failures occurred at the terminal-step transition; tread slip was the recorded failure mode in all three cases, and the models attribute it to a combination of insufficient drive-torque margin (≈ 83 N·m demand per motor versus a 45 N·m continuous and an 85 N·m peak rating) and marginal tread traction (required friction coefficient 0.64 versus ≈ 0.65 available). For the 100 kg design payload (130 kg total mass) the models predict ≈ 106 N·m per motor, above the peak rating, and an actuator safety factor of only 1.33, and the preliminary static-stability estimate indicates that the chair is not self-stable on the flight without attendant support. The measured actuator extension (92 ± 3 mm) and the current-based peak torque estimate (83 N·m) were consistent with the model predictions within 5%. The IoT layer achieved a mean remote-command latency of 180 ms, an obstacle-stop latency of 45 ms and a GPS accuracy of 2.5–3.5 m, at a prototype component cost of USD 350–450. The results demonstrate the feasibility of the low-cost architecture but not yet its reliability for users; specific design corrections and a validation plan are proposed. No human participants were involved, and the findings do not constitute clinical validation.