Cycling by means of functional electrical stimulation (FES) is an attractive training method for individuals with paraplegia. The physiological benefits of FES are combined with the psychological incentive of independent locomotion. In addition, cycling has the advantage in that the generated muscle forces are converted into drive power with relatively high efficiency compared to other means of locomotion, e.g., walking. For the design of an appropriate cycling device and the development of optimal stimulation patterns, it has to be investigated how the geometry for FES cycling, influenced by individual parameters of the FES-generated drive torques and the magnitude of variations among subjects with paraplegia, can be optimized. This study shows the design of a freely adjustable test bed with additional motor drive which allows static and dynamic measurements of force components and drive torque at the crank. Furthermore, the influence of geometry and various individual parameters on FES pedaling can be tested for each subject individually. A pedal path realized by a three-bar linkage that was optimized according to preliminary simulations further increases leg cycling efficiency. Safety precautions avoid injuries in case of excessive forces, e.g., spasms. Test results illustrate the application of the test bed and measurement routines. A test series with four paraplegic test persons showed that the presented static and dynamic measurement routines allow to provide optimal stimulation patterns for individual paraplegic subjects. While pedaling with these optimal stimulation patterns only negligible negative active drive torques, due to active muscle forces, were applied to the crank and sufficient drive power was generated to power a cycle independently.

The key to robust and yet accurate capacitive angular-position sensors is, besides a smart measurement principle, the design of the charge amplifier. In a given context robust means insensitivity to dirt, moisture and dew on the receiving electrodes. Other environmental influences play a secondary role and can be tackled routinely. This paper presents a charge amplifier model, and compares simulation results with measurements obtained by operating the amplifier under harsh environmental conditions. To verify these simulation results and to distinguish between measurement errors originating from the sensor topology and from the charge amplifier, a measurement system i s demonstrated to observe charge amplifier in fluences on the measurement accuracy in isolation.With the aid of this system and the corresponding model, an improved charge amplifier circuit has been found. The achieved accuracy is fairly high even under the influence of large leakage currents. The superiority of the present model is compared to conventional designs.