Capsule Neural Network based Height Classification using Low-Cost Automotive Ultrasonic Sensors
This work addresses the need for reliable object height analysis in autonomous driving scenarios using cost-effective hardware, representing an incremental improvement in sensor data processing.
The paper tackled the problem of performing detailed height classification of objects using low-cost automotive ultrasonic sensors by proposing a capsule neural network approach with a novel resorting and reshaping method, achieving outstanding results in classification performance and computation speed.
High performance ultrasonic sensor hardware is mainly used in medical applications. Although, the development in automotive scenarios is towards autonomous driving, the ultrasonic sensor hardware still stays low-cost and low-performance, respectively. To overcome the strict hardware limitations, we propose to use capsule neural networks. By the high classification capability of this network architecture, we can achieve outstanding results for performing a detailed height analysis of detected objects. We apply a novel resorting and reshaping method to feed the neural network with ultrasonic data. This increases classification performance and computation speed. We tested the approach under different environmental conditions to verify that the proposed method is working independent of external parameters that is needed for autonomous driving.