Accuracy limits of in-room localisation using RSSI

An ML approach is able to exploit all available information, and has proven to be a powerful method to evaluate RSSI localisation methods. We plan to use it to attack problems like optimal anchor placement, exploiting multiple mobile receivers, and localisation coupled to tracking. One preliminary result is that, when accounting for positioning and RSSI errors in a specific single-room scenario, we needed at least 12 anchors to obtain a third quartile localisation error of less than 150 cm. This suggests that much investigation is needed for single-room RSSI localisation to attain any useful performance.