IJSER © 2019 http://www.ijser.org An Intelligent System for Automatic Footdrop Correction in Stroke Patients using FES: A Pilot Study Naveen Gangadharan, Sivakumar Balasubramanian, George Tharion, Judy John, Thangavelu Senthilvelkumar, Suresh Devasahayam Abstract— Initial post-stroke hemiparesis is common in stroke patients that might lead to motor impairments of the contralateral limbs. They usually are presented with impaired ankle-foot function, commonly termed as footdrop / dropfoot. Their gait is associated with foot slap, toe-drag and hip-circumduction. Despite the research in post-stroke rehabilitation that has brought in various technical insights to footdrop correction, there is still want of evidence-based rehabilitation guidelines because of limited understanding of the mechanisms leading to footdrop and its correction. This paper presents a study with main objectives being 1) to develop a low-cost inertial motion sensor-based footdrop correction system that uses Functional Electrical Stimulation (FES) as intervention, and to assess the effectiveness of the system in correcting footdrop with manual stimula- tion using a press button, and 2) propose an algorithm based on the above results to automate the stimulation timing of the device. Six healthy subjects and two stroke survivors were recruited for the study. Studies related to FES-based footdrop correction has always presented problems pertaining to an efficient way of achieving a normal gait. There are lesser evidences on the parameters of stimulation including the timing of stimu- lation which is a prime factor to achieve a smooth normal gait pattern which this study has taken into consideration. The results of this study show that such a device is expected to help stroke survivors with footdrop to walk with enough clearance. The tibial tilt angle, tibial angular velocity and forefoot normal acceleration components have been used to simulate the automatic stimulation ON/OFF pulse and the algorithm is found to work for the controls recruited, which proves the feasibility of automating the stimulation using sensor-based swing phase detection. The positive feed- back about the device has also shown a direction towards the future work this work demands to completely automate the system and make a reli- able, low-cost, user friendly footdrop correction device. Index Terms— FES (Functional Electrical Stimulation), footdrop, gait analysis, IMU (Inertial Motion Sensor), Madgwick algorithm, peroneal stimulator, stroke rehabilitation, tibialis anterior stimulation, walkaide. . —————————— ◆ —————————— 1 INTRODUCTION Stroke survivors are often presented with contralateral hem- iplegia if there is damage to the corticospinal tract and there will be persistent distal weakness, one such case being dropfoot. The subject will not be able to actively dorsiflex the foot during the swing phase of the gait. Hip circumduction, steppage gait, toe dragging, reduced gait speed, and higher fall risks are usually presented as the consequences. As com- pared to the conventional AFOs that limits ankle mobility leading to contractures, discomfort and unfavourable aes- thetics, FES has emerged as one of the effective means of achieving active ankle dorsiflexion during the swing face, helping the stroke survivors to attain good foot clearance and achieve a natural gait [7] [8] [9]. A peroneal nerve stimu- lator can be used to stimulate the common peroneal nerve innervating the tibialis anterior responsible for the ankle dor- siflexion [10] [1]. The peroneal stimulator design has been in research for decades to develop a more user friendly, light weight, more reliable, and economical orthotic device. The early 1990s wit- nessed use of surface electrodes and foot switches (either open/ close mechanical switch or force sensitive resistors) with limitations of being required to be worn along with a stable foot wear or some other means holding the switch and the major drawback was the inappropriate firing of the switch due to poor contact during the hemiplegic gait. The entire set up was not convenient to use on daily basis. To overcome these limitations, sensors with accelerometers and gyroscopes, that can detect the joint orientation and the timing of stimulation for dorsiflexion can be used to calculate gait kinematics obtained from the sensor data [11] [12]. This presented study is on the development of a sensor-based footdrop correction device using FES with manual mode of triggering stimulation, being tested on normal volunteers and stroke patients. The data collected has been used to pro- pose an algorithm for automatic footdrop correction where the stimulation timing is decided based on the manual stimulation data [16]. This would be the first step towards the design and development of a user friendly, motion sensor-based dropfoot correction system with FES device, having a manual control and an automatic control. The commercial available model of this kind of a footdrop correction system has been reviewed by the users with the following drawbacks: i) the device stimulation timing accuracy does not take into account the situations where a user sits by bending the knees and the device wrongly stimulates detecting the tibia tilt, ii) there is no provision to switch back to the manual control mode from the automatic mode if the user decides to switch. This study would be an attempt to check the feasibility of such a device adapted to Indian settings in terms of cost, provision for barefoot wearing and accuracy. 2 METHODS The apparatus comprised of a multichannel stimulator that International Journal of Scientific & Engineering Research Volume 10, Issue 4, April-2019 ISSN 2229-5518 1,102