Functional Electrical Stimulation (FES) systems for gait rehabilitation after stroke commonly rely on open-loop control strategies based on foot switches, with their main stimulation target being the tibialis anterior muscle, innervated by the peroneal nerve. The desired effect in such systems is assistance of the ankle dorsiflexion (foot lift) during the swing phase of gait. However, foot switches in FES systems present shortcomings regarding ergonomics, durability and personalization. As an alternative to foot switches, inertial measurement units (IMUs) have been used to detect gait events as part of FES control strategies. However, in many stroke patients not only ankle dorsiflexion, but also plantar flexion is impaired. Few works have reported the use of IMUs to detect gait events for assistance of both key ankle movements. The main aim of this project was to study the use of IMUs for the detection, in real time, of two relevant gait events: heel strike and toe off. A secondary aim was to explore the feasibility of using the proposed detection method as a part of a closed loop system that controls two FES channels (eventually targeted to tibialis anterior and gastrocnemius muscles). This proposal is based on previous offline movement analysis studies that have shown the usefulness of IMUs to define pathological and non-pathological gait patterns. A comparison was made between the data collected from two IMUs placed at two different points on the lower limb and the data obtained was used to develop an algorithm to detect two gait events: Initial (IC) and End Contact (EC). A Python-based script was developed and executed on a single board computer (Raspberry Pi 4), to simultaneously read sensor information, run the detection algorithm, and deliver, in real-time, control commands to two electrical stimulation channels, as dictated by a Finite State Machine (FSM) controller. A healthy volunteer was instrumented with the system and asked to walk, while the acquisition system and detection algorithm were operating. 100% correct detections of IC (13/13) and EC (10/10) events were obtained, without considering the steps needed to change directions. Moreover, it was verified with an oscilloscope that the correct stimulation sequences were delivered by the FES device, in accordance with the events detected and the FSM conditions. Additionally, it was found that the vertical axis of the angular velocity information derived from a single IMU placed on the sagittal plane of the lower limb has the potential to provide reliable information for the detection of gait events using the proposed algorithm, which would avoid the use of the two IMU sensors used in this work. With the system proposed in this project, a first approach has been achieved to- wards real-time detection of gait events and its feasibility for use as a closed-loop control system for an FES system. This system has the potential to acquire robustly acceleration and angular velocity data from both pathological and healthy gait in different populations, in rehabilitation settings, for movement assessment and rehabilitation applications, when coupled with FES or other type of actuators.