Many techniques for robot localization rely on the assumption that both process and measurement noises are uncorrelated, white, and normally distributed. However, if this assumption does not hold, these techniques are no longer optimal and, in addition, the maximum estimation errors can be hardly kept under control. In this paper, this problem is addressed by means of a tailored extended H $_\infty$ filter (EHF) fusing odometry and gyroscope data with position and heading measurements based on quick response (QR) code landmark recognition. In particular, it is shown that, by properly tuning EHF parameters and by using an adaptive mechanism to avoid finite escape time phenomena, it is possible to achieve a higher localization accuracy than using other dynamic estimators, even if QR codes are detected sporadically. Also, the proposed approach ensures a good tradeoff in terms of computational burden, convergence time, and deployment complexity.

This paper presents the application of the Synthetic Aperture Radar (SAR) localization method for indoor positioning of UHF-RFID tags when the robot-mounted reader antenna moves along multiple trajectories. By properly combining the phase data associated to a set of multiple paths, the whole length of the combined synthetic apertures enlarges and then the localization accuracy may improve. Besides, during consecutive inventory rounds, several tag position estimates are available and they can be profitably combined to minimize the localization uncertainty. Different combination approaches are investigated to determine the best choice to improve the localization performance. The method capabilities are discussed through a numerical analysis, by considering different configurations of the multiple apertures and different measurement uncertainty sources. Finally, the proposed localization method is validated through an experimental analysis carried out with commercial RFID hardware and a robotic wheeled walker, in an indoor scenario, by employing different types of tags. The knowledge of the reader/robot trajectory required by the SAR method is here achieved with an optical system.

The ultra-wide band radio technology (UWB) is currently considered the de-facto standard for implementing precise indoor positioning systems. Several positioning algorithms are currently being investigated for finding the best implementation in terms of scalability, refresh rate, and energy requirements. Among all the proposed approaches, the Downlink Time Difference of Arrival (DTDoA) is currently considered one of the most promising techniques capable of tracking any number of assets without decreasing the measurement update rate. This paper proposes a model for the DTDoA and validates the algorithms on UWB data, using a motion capture system as ground truth. Results highlight the validity of the previously proposed model, showing that the proposed UWB indoor positioning system achieves a maximum 30 cm uncertainty with only a simple wireless synchronisation, avoiding wired procedure that limits the usability of the infrastructure.

The phase of signals backscattered by Ultra High Frequency (UHF) Radio Frequency Identification (RFID) tags is generally more insensitive to multipath propagation than Received Signal Strength Indicator (RSSI). However, signal phase measurements are inherently ambiguous and could be further affected by unknown phase offsets added by the transponders. As a result, the localisation of an agent by using only signal phase measurements looks infeasible. In this paper, it is shown instead that the design of a dynamic position estimator (e.g. a Kalman filter) based only on signal phase measurement is actually possible. To this end, the necessary conditions to ensure theoretical local nonlinear observability are firstly demonstrated. However, a system that is locally observable guarantees convergence of the localisation algorithm only if the actual initial agent position is approximately known a priori. Therefore, the second part of the analysis covers the global observability, which ensures convergence starting from any initial condition in the state space. It is important to emphasise that complete observability holds only in theory. In fact, measurement uncertainty may greatly affect position estimation convergence. The validity of the analysis and the practicality of this localisation approach are further confirmed by numerical simulations based on an Unscented Kalman Filter (UKF).

We consider a vehicle consisting of a robotic walking assistant pushed by a user. The robot can guide the person along a path and suggest a velocity by various means. The vehicle moves in a crowded environment and can detect other pedestrians in the surroundings. We propose a reactive planner that modifies the path in order to avoid pedestrians in the surroundings. The algorithm relies on a very accurate model to predict the motion of each pedestrian, i.e., the headed social force model. The possible trajectories for both the vehicle and the pedestrians are modeled as clothoid curves, which are efficient to manage from the numeric point of view and are very comfortable to follow for the user. Probabilistic techniques are used to account for the variability of the motion of each pedestrian. The path is efficient to generate, is collision free (up to a certain probability), and is comfortable to follow. Simulations and comparisons with a state-of-the-art planner using real data as well as experiments are reported to prove the effectiveness of the method.