Indoor positioning system based on visible light communication and location fingerprinting can achieve cm-level positioning accuracy, which has become an important candidate for high precision positioning. However, in the case of non-uniform and low-density location fingerprinting distribution, it will still lead to large errors and poor robustness. To solve this problem, an accurate visible light positioning (VLP) method based on location fingerprinting and meta-heuristic is proposed, and all possible cases of uneven distribution of location fingerprinting are considered by introducing the random selection rate. The effectiveness of the proposed VLP method is compared with other six VLP methods based on location fingerprinting. Simulation results show that when the signal-to-noise ratio is 20 dB, the average positioning error of the proposed VLP method is 0.51 cm, which is at least 80.23% lower than the other six VLP methods. Experimental results show that when the random selection rate is 50%, the average positioning error of the proposed VLP method is 3.11 cm, which is at least 41.87% lower than the other six VLP methods. The proposed VLP method can be applied to positioning scenarios with different location fingerprinting distributions. The research results can provide a new research idea for the fingerprinting weight calculation method.

The weighted K-nearest neighbor (WKNN) algorithm is a commonly used fingerprint positioning, the difficulty of which lies in how to optimize the value of K to obtain the minimum positioning error. In this paper, we propose an adaptive residual weighted K-nearest neighbor (ARWKNN) fingerprint positioning algorithm based on visible light communication. Firstly, the target matches the fingerprints according to the received signal strength indication (RSSI) vector. Secondly, K is a dynamic value according to the matched RSSI residual. Simulation results show the ARWKNN algorithm presents a reduced average positioning error when compared with random forest (81.82%), extreme learning machine (83.93%), artificial neural network (86.06%), grid-independent least square (60.15%), self-adaptive WKNN (43.84%), WKNN (47.81%), and KNN (73.36%). These results were obtained when the signal-to-noise ratio was set to 20 dB, and Manhattan distance was used in a two-dimensional (2-D) space. The ARWKNN algorithm based on Clark distance and minimum maximum distance metrics produces the minimum average positioning error in 2-D and 3-D, respectively. Compared with self-adaptive WKNN (SAWKNN), WKNN and KNN algorithms, the ARWKNN algorithm achieves a significant reduction in the average positioning error while maintaining similar algorithm complexity.

This article proposes an indoor cooperative positioning system for a swarm of micro unmanned aerial vehicles (UAVs) based on visible light communication. The proposed cooperative positioning system is composed of light-emitting diodes (LEDs) installed on the ceiling and a group of micro UAVs. Each micro UAV is composed of a single LED and multiple photodiodes (PDs) with a uniform distribution. The cooperative positioning of a group of micro UAVs is achieved through the LEDs installed on the ceiling and the interactive positioning information shared among them. In order to solve the nonconvex optimization function constructed by cooperative localization, an improved bat algorithm based on adaptive parameters was designed. In the proposed cooperative localization scenario, the theoretical lower bound of the localization error, i.e., the Cramer–Rao lower bound, is precisely analyzed by mathematical method. Simulation and experimental results demonstrate that compared with the single PD receiving technology, the multi-PD receiving technology can significantly improve the positioning accuracy, while the cooperative positioning technology can significantly improve the positioning accuracy compared with the noncooperative positioning technology.

Although the fingerprint-based visible light positioning (VLP) method can achieve centimeter-level accuracy, it requires collecting received signal strength (RSS) values from a large number of training locations in the offline stage, especially when considering the rotation angle of the photodiode (PD). To enhance the practicality of VLP based on the fingerprint method, this paper proposes a novel fingerprint database regeneration method. The proposed method does not rely on a signal propagation model but only requires knowledge of the actual coordinates of LEDs while also considering the impact from the rotation angle of the PD, making it more practical. Simulation and experimental results demonstrate that the proposed method achieves the highest positioning accuracy compared to the other four fingerprint database regeneration methods. The proposed method can still achieve satisfactory localization performance even if the proportion of training locations is low and randomly distributed. When the number of training fingerprints is 60 and randomly distributed, the proposed method can still achieve localization performance similar to that of an actual dense fingerprint database. Even when considering the rotation angle of the PD, the proposed method reduces the average positioning error by 39.85% in simulation and 29.78% in experiment, respectively, compared to sparse fingerprint databases.

This work investigates the problem of simultaneously positioning and orientating (SPAO) the user equipment (UE) with reconfigurable intelligent surface (RIS)-assisted visible light communications (VLCs). In the proposed system model, the utilization of RISs effectively helps to address the SPAO problem of the UE in situations when there are insufficient light-emitting diode (LED) transmitters. Firstly, the channel gain between LED and UE and the one between RIS and UE are obtained through signal separation, followed by formulating an objective optimization function for the SPAO problem. A closed-form Cramer-Rao lower bound (CRLB) is then derived, representing the theoretical lower limit of positioning and orientating errors. Consequently, using CRLB as a benchmark for the SPAO performance, a fuzzy distance-based particle swarm optimization (FDPSO) is proposed to minimize the deviations between SPAO estimation errors and CRLBs. The proposed FDPSO algorithm adaptively adjusts the parameters and mutation operations to balance local exploitation and global exploration. Simulation results demonstrate that the proposed FDPSO algorithm achieves an asymptotic tightness with CRLB and performs better than the other five related iterative algorithms in terms of the cumulative distribution functions (CDFs) of positioning errors (PEs) and orientating errors (OEs).