A comparative analysis between deep neural network-based 1D-CNN and LSTM models to harness the self-sensing property of the Shape Memory Alloy wire actuator for position estimation

The article presents a performance based comparative analysis of popular deep neu- ral network(DNN) models such as 1D-CNN and LSTM for Shape Memory alloy(SMA)- based wire actuator position estimation. These DNN models utilize the self-sensing property(SSP) for position prediction of the SMA actuator. The phase dependent elec- trical resistivity of SMA wire act as SSP, where the electrical resistivity is in the form of resistance acts as inputs to the proposed models for precise estimation of current position of the SMA actuator. For effective position control of the actuator, an ac- curate sensor feedback is required, utilizing SSP results in the elimination of external position sensor. This will improve the overall system in terms of more compactness and reduced interface complexity. Coming to DNN models, 1-D CNN has been meagerly explored in the current literature landscape for self-sensing prediction of SMA actua- tors, these 1-D CNN models are becoming quite popular for time series prediction for various applications and are emerging as an alternative to LSTM models. In this paper, a novel implementation of a 1D-CNN model for SMA actuator position estimation has been done. A comparative analysis between 1D-CNN and LSTM has been done for prediction capability and inference speed based on performance measures such as Mean Square Error(MSE), Mean Absolute Error(MAE), sMAPE(symmetric Mean Absolute Percentage Error), data distribution and average inference speed. This comparison shows that 1D-CNN has matching performance with the LSTM model with respect to the prediction capability, however 1D-CNN offers faster inference speed. This analy- sis can be useful for choosing suitable DNN model for deployment in low computing platform such as micro-controller for SMA actuator based real time applications where time latency is a critical parameter.