ABSTRACT
Intelligent field robotic systems are mobile robots that operate in an unconstrained environment that is constantly changing with regard to both internal and external states. Learning, reasoning or adaptive behaviour generation is what differentiates mobile field robotic systems from its industrial counterparts; learning is not required when the environment is constrained to the extent that pre-programming is sufficient for the robot to achieve its objectives. Mobile robots rely on sensors to generate a description of the internal states and external environment. Data fusion in mobile robotics occur at different levels; there is time-series data fusion of a single sensor, fusion of data from redundant sensors, multiple sensor data fusion and fusion of sensor and goal information. Although research into data fusion is advancing at a steady rate, challenges to accurate data fusion for mobile, intelligent and autonomous systems remains predominantly the characterization of the uncertainty in sensor measurements.
