Learning-Based Bias Correction for Time Difference of Arrival Ultra-Wideband Localization of Resource-Constrained Mobile Robots

Accurate indoor localization is a crucial enabling technology for many robotics applications, from warehouse management to monitoring tasks. Ultra-wideband (UWB) time difference of arrival (TDOA)-based promising lightweight, low-cost solution that can scale large number devices-making it especially suited resource-constrained multi-robot applications. However, the accuracy standard, commercially available UWB radios often insufficient due significant measurement bias and outliers. In this letter, we address these issues by proposing robust TDOA framework comprising (i) learning-based correction (ii) M-estimation-based filtering handle The key properties our approach are learned biases generalize different anchor setups computationally efficient enough run on hardware. We demonstrate Crazyflie nano-quadcopter. Experimental results show proposed framework, relying only onboard IMU UWB, provides an average 42.08% error reduction (in three setups) compared baseline without compensation. also autonomous trajectory tracking quadcopter using approach.