Hardware-Accelerated Data Processing of Capacitive Sensor Arrays for Industrial and Service Robotic Applications

Abstract:
To ensure a safe environment for human-machine interaction through signal processing and to process big data from intelligent production systems, increasingly powerful computational hardware is required. Central Processing Unit (CPU) based systems are reaching their load limits when processing large amounts of data. This paper presents a hardware-software architecture for Field Programmable Gate Array (FPGA) based platforms for high-frequency data evaluation using capacitive sensor arrays as an example. The architecture includes on the one hand a concept for hardware-acceler-ated sensor data processing and on the other hand the methodology for reading out capacitive sensor arrays. The focus is on simple and easily scalable electronics as well as a dynamic data processing module that can execute and deterministically process complex algorithms with large amount of data based on flexible rules using hardware acceleration. The presented concept was implemented on a FPGA based hardware platform. This allowed the signal processing of 100 sensor values with a resolution of 32 Bit each by 3 pipelined rules, a second order IIR-filter, a 10-times linear interpolation and an averaging algorithm at a task cycle of 1kHz. For the capacitance measurement validation, a custom developed Printed Circuit Board (PCB) with 6 channels is connected to the FPGA platform. It was shown to be capable of a robust and deterministic capacitance measurement at intervals of 1ms with capacitance values from Femto- to Nanofarad. Due to its easy scalability, the method can be used, for example, to completely cover a robotic arm similar to human skin, fully evaluate its measurement signals and therefore increases Human Machine Interaction (HMI).