The inverted pendulum is a classic problem in dynamics and control theory due to its inherently unstable nature. In the system tested, Field Programmable Gate Arrays (FPGAs) are used for the implementation of control and sensor fusion algorithms in the inertial navigation system of a Mobile Inverted Pendulum (MIP) robot. Additionally, the performance of digital PID control and Kalman filter algorithms are tested in this FPGA system. The test platform for tuning Kalman filter is designed using

more »

... tical encoders as a standard reference. PWM signal generation and quadrature phase decoding of encoder pulses is accomplished using hardware description language in FPGA. The values from the inertial sensors and quadrature phase decoded values are fed into MicroBlaze, a 32-bit soft-core RISC processor, within the FPGA. The overall system demonstrates the use of low cost inertial sensors to balance a two wheeled robot. The system is presently able to balance on its own and it also serves as an extremely reconfigurable FPGA based platform to facilitate future modifications, updates and enhancements with more complex control and sensor fusion techniques. The robot framework (Figure 1) , 50 cm in height, is comprised of adjustable plywood stacks used for placement of electronic components. Two motion sensors, an accelerometer and a gyroscope, are used to continuously calculate, via dead reckoning, the orientation of the MIP robot without the need for external references. These two sensors are low cost but their measurements are corrupted with noise. The orientation of the MIP robot must be sensed with higher precision to