|Improved Haptic Fidelity via Reduced Sampling Period with an FPGA-Based Real-Time Hardware Platform
|Year of Publication
|O'Malley, MK, Sevcik, KS, Kopp, E
|ASME Journal of Computing and Information Science in Engineering
A haptic virtual environment is considered to be high-fidelity when the environment is perceived by the user to be realistic. For environments featuring rigid objects, perception of a high degree of realism often occurs when the free space of the simulated environment feels free and when surfaces intended to be rigid are perceived as such. Because virtual surfaces (often called virtual walls) are typically modeled as simple unilateral springs, the rigidity of the virtual surface depends on the stiffness of the spring model. For impedance-based haptic interfaces, the stiffness of the virtual surface is limited by the damping and friction inherent in the device, the sampling rate of the control loop, and the quantization of sensor data. If stiffnesses greater than the limit for a particular device are exceeded, the interaction between the human user and the virtual surface via the haptic device becomes nonpassive. We propose a computational platform that increases the sampling rate of the system, thereby increasing the maximum achievable virtual surface stiffness, and subsequently the fidelity of the rendered virtual surfaces. We describe the modification of a PHANToM Premium 1.0 commercial haptic interface to enable computation by a real-time operating system (RTOS) that utilizes a field programmable gate array (FPGA) for data acquisition between the haptic interface hardware and computer. Furthermore, we explore the performance of the FPGA serving as a standalone system for communication and computation. The RTOS system enables a sampling rate for the PHANToM that is 20 times greater than that achieved using the “out of the box” commercial hardware system, increasing the maximum achievable surface stiffness twofold. The FPGA platform enables sampling rates of up to 400 times greater, and stiffnesses over 6 times greater than those achieved with the commercial system. The proposed computational platforms will enable faster sampling rates for any haptic device, thereby improving the fidelity of virtual environments.