The lower limb movement intentions of stroke patients are of great significance for assisting training with lower limb rehabilitation robots.This paper proposes a low-cost,structurally simple and high-load-bearing multi-dimensional foot force sensor system designed for the recognition of movement intentions in lower limb rehabilitation robots.Firstly,a detailed description of the design and analysis of the multi-dimensional foot force sensor system is provided,including foot force analysis,system structure design,and the establishment of a force/torque conversion model.Subsequently,through experiments,the performance of the sensor system platform is validated.The results demonstrate that the designed sensor system can stably and accurately detect changes in the foot forces of the patients,providing reliable data support for precise recognition of movement intentions.This method holds promise for applications in lower limb rehabilitation training for stroke patients,thereby enhancing the effectiveness of rehabilitation.