Bayesian and Neural Inference on LSTM-based Object Recognition from Tactile and Kinesthetic Information
This work addresses object recognition for robotic manipulation, but it is incremental as it applies existing methods like LSTMs and Bayesian fusion to a specific domain.
The paper tackled multimodal object recognition by fusing tactile and kinesthetic data using LSTM networks and Bayesian/Neural inference, showing that Bayesian-based classifiers improved recognition capabilities and outperformed the Neural-based approach in a 36-class experiment.
Recent advances in the field of intelligent robotic manipulation pursue providing robotic hands with touch sensitivity. Haptic perception encompasses the sensing modalities encountered in the sense of touch (e.g., tactile and kinesthetic sensations). This letter focuses on multimodal object recognition and proposes analytical and data-driven methodologies to fuse tactile- and kinesthetic-based classification results. The procedure is as follows: a three-finger actuated gripper with an integrated high-resolution tactile sensor performs squeeze-and-release Exploratory Procedures (EPs). The tactile images and kinesthetic information acquired using angular sensors on the finger joints constitute the time-series datasets of interest. Each temporal dataset is fed to a Long Short-term Memory (LSTM) Neural Network, which is trained to classify in-hand objects. The LSTMs provide an estimation of the posterior probability of each object given the corresponding measurements, which after fusion allows to estimate the object through Bayesian and Neural inference approaches. An experiment with 36-classes is carried out to evaluate and compare the performance of the fused, tactile, and kinesthetic perception systems.The results show that the Bayesian-based classifiers improves capabilities for object recognition and outperforms the Neural-based approach.