FES for restoring terrain-specific human locomotion

FES for restoring terrain-specific human locomotion 1st Andrea Demofonti CREO Lab Universit` a Campus Bio-Medico di Roma Rome, Italy [email protected] 2nd Marica Ferrara CREO Lab Universit` a Campus Bio-Medico di Roma Rome, Italy [email protected] 3rd Loredana Zollo CREO Lab Universit` a Campus Bio-Medico di Roma Rome, Italy [email protected] 4th Francesca Cordella CREO Lab Universit` a Campus Bio-Medico di Roma Rome, Italy [email protected] Abstract—Functional Electrical Stimulation (FES) can assist patients walking since it induces muscle contractions through the non-invasive delivery of electrical pulses. Literature studies have shown that environmental factors affect patients’ myoelectric activity during gait. However, no closed-loop FES system has yet integrated such information to modulate stimulation. The aim of this study is to develop a novel closed-loop multichannel FES system for lower limb assistance, which adapt stimulation on the basis of both exteroceptive and proprioceptive information. The system performance was evaluated on seven leg muscles of five healthy participants during walking in five different scenarios: Up Ramp (UR), Up Stair (US), Level Ground (LG), Down Ramp (DR) and Down Stair (DS). The system can classify terrains with 98% accuracy and successfully activate muscles achieving success rates of 100% except for US (90%) and DS (93%). Index Terms—Functional Electrical Stimulation (FES), human locomotion, lower limb assistance, surface electromyography (EMG), terrain recognition. I. INTRODUCTION LOWER limb sensorimotor disabilities may compromise patients’ capabilities during locomotion. In this field, Functional Electrical Stimulation (FES) can enhance motor function since it provides electrical pulses in a non invasive way to generate muscular contractions [1]. Most closed-loop FES systems rely on patient muscular residual capabilities to provide an on-line stimulation modulation [2]. Nonetheless, it was demonstrated how the terrain affects the lower limb muscles activation patterns during walking [3] but systems capable of modulating stimulation according to both the patient’s proprioceptive information and exteroceptive one have not been proposed in literature. Therefore, this study aimed to develop and test a novel closed-loop multichannel FES system for lower limb assistance capable of adapting stimulation according to both exteroceptive and proprioceptive information. The proposed system was tested on five healthy participants in five different walking scenarios. II. MATERIALS AND METHODS The proposed system was capable of adapting stimulation according to both exteroceptive and proprioceptive information (Fig.1). The first one was on-line evaluated thanks to a Terrain Recognition Optical System (TROS) fastened at the This study was supported by the Italian Institute for Labour Accidents (INAIL) Prosthetic Center with the BioARMnext (CUP: E58D19000650005) and BioInterNect (CUP: E57G23000280005) projects. waist. Once the terrain was discriminated, the corresponding myoelectric patterns were selected to off-line modulate FES. Fig. 1. Working principle of the proposed closed-loop FES system. The TROS is composed of the infrared distance sensor GP2Y0A02YK0F and an electronic board for data acquisition. The sensor and the board were fixed to a 3D-printed support fastened to the participant’s waist through velcro straps with an inclination angle of 0.78 rad. The adopted terrain recognition algorithm was an optimized version of a previously reported approach [4] and it was characterized by a two-layer decision tree. The first one provided a course discrimination (Upper Terrain, UT, Level Ground, LG, Lower Terrain, LT), while the second one delivered a more refined output classifying uneven terrain into the following additional classes: Up Ramp (UR), Up Stair (US), Down Ramp (DR), Down Stair (DS) and Others (OT) [4]. The symmetric biphasic square wave was adopted [5]–[8]: the Pulse Amplitude (PA) and the Pulse Frequency (PF) were kept constant to PAMAX (i.e., a current value specific for each muscle of each participant) and 40 Hz [1], respectively; the PW was varied as follows PW(V) = (P WM AX −P WM IN )(V−V0) VM AX −V0 +P WM IN (1) where PWMAX and PWMIN were the maximum and minimum PW value specific for each muscle of each participant, respectively. V, V0and VMAX indicated the current, minimum and maximum value of the normalized and enveloped EMG signal of the muscle of interest. Five healthy participants 2025 I-RIM Conference October 17-19, Rome, Italy ISBN: 9788894580570 10.5281/zenodo.17629858 219 (3M/2F, 22.8±2.7 years) were enrolled. The experimental setup included: i) the TROS; ii) seven EMG Trigno Wireless sensors for monitoring the electrical activity of the following muscles of the participants’ right leg: 1) Rectus Femoris (RF); 2) Vastus Lateralis (VL); 3) Biceps Femoris (BF); 4) Semitendinosus (ST); 5) Tibialis Anterior (TA); 6) Gastrocnemius Medialis (GM); 7) Soleus (SO); iii) the multichannel electric stimulator STG4008 for delivering stimuli through two circular (50 mm diameter), auto-adhesive and superficial electrodes located over each muscle. The experimental protocol was divided into three phases. The first one aimed at evaluating the electrical activity of the participants’ right leg muscles to be used as reference for the FES modulation. The participants were equipped with EMG sensors and walked in a random order across four scenarios (3 repetitions each): LG–UR, DR–LG, LG–US and DS–LG. The second phase aimed to the identification of the FES electrodes’ optimal position for each muscle of the participants’ right leg and the relative maximum/minimum stimulation parameters. The third phase aimed to the evaluation of the participants’ deambulation performance with the proposed FES system. Therefore, the participants performed the previous activities via FES and therefore without any volitive muscular contraction (see Fig.2). The performance of the proposed system was quantified in terms of: i) PAMAX values; ii) the TROS accuracy in terrain recognition; iii) the system efficacy: a score of 1 or 0 was assigned to each repetition depending on whether the participants were able to carry out them successfully or not. Hence, a Success Rate (SR) was introduced as the percentage of successfully completed activities compared to the total. Fig. 2. ALateral view of a representative participant. B-FRepresentative participant using the proposed system in the five different walking scenarios. III. RESULTS AND DISCUSSION The median values of PAMAX of each muscle are: 15.5 mA for VL, 13.5 mA for both RF and SO, 15.0 mA for both BF and ST, 14.5 mA for GM and 13.0 mA for TA (no statistically significant differences, Wilcoxon signed-rank test with Bonferroni correction, P<0.0024). As far as the terrain coarse discrimination, the system correctly recognized the UT, LG and LT respectively in 98%, 100% and 99% of the time (see Fig.3). As far as the terrain fine discrimination, the system correctly discriminated the UR, US, LG, DS and DR respectively in 97%, 100%, 98%, 98% and 97% of the time (see Fig.3). The FES correctly assisted participants during deambulation obtaining SR values of 100% in all the walking scenario except for US (90%) and DS (93%). The reduced SR values observed in these conditions is likely attributable to the increased kinematic complexity due to stairs presence. Fig. 3. Normalized confusion matrices of coarse (A) and fine (B) terrain recognition. IV. CONCLUSION In this study, a novel closed-loop multichannel FES system for lower limb assistance adapting stimulation according to both exteroceptive and proprioceptive information was designed, developed and evaluated on five participants for walking in five different scenarios. The results demonstrated that the system can classify terrains with 98% accuracy and successfully activate muscles to perform the tasks obtaining SR values of 100% for UR, LG, and DR, 90% for US and 93% for DS. Further studies will be dedicated to an experimental validation of the proposed system on post-stroke patients evaluating its efficacy, usability and comfort. REFERENCES [1] Lynch, C. L., & Popovic, M. R. (2008). Functional electrical stimulation. IEEE control systems magazine [2] Melo, P. L. et al. (2015). Technical developments of functional electrical stimulation to correct drop foot: sensing, actuation and control strategies. Clinical Biomechanics [3] Scherpereel, K. et al. (2023). A human lower-limb biomechanics and wearable sensors dataset during cyclic and non-cyclic activities. Scientific Data [4] Liu, M. et al. (2015). Development of an environment-aware locomotion mode recognition system for powered lower limb prostheses. IEEE Transactions on Neural Systems and Rehabilitation Engineering [5] Collu, R. et al. (2023). Wearable high voltage compliant current stimulator for restoring sensory feedback. Micromachines [6] Mereu, F. et al. (2024). A Sensory Feedback Neural Stimulator Prototype for Both Implantable and Wearable Applications. Micromachines [7] Paolini, R. et al. (2025). A novel portable device and validation procedure for transcutaneous electrical nerve stimulation. Wearable Technologies [8] Paolini, R. et al. (2025). Wearable stimulator for upper and lower limb somatotopic sensory feedback restoration. IEEE Transactions on Biomedical Circuits and Systems. 220