Adaptive gait detection based on foot-mounted inertial sensors and multi-sensor fusion

•Dual-gait detection to utilize the coupling relationship between lower limbs.•A rule-based detection method to yield enough training data.•A 6-state left-to-right hidden Markov (HMM) model to cover more gait details.•A neural network (NN) to handle the high dimensional gait data and feed the HMM.•The HMM corrects misclassifications of the NN by providing contextual information.