A camera assisted received signal strength (CA-RSS) algorithm is proposed for visible light positioning (VLP) systems. The basic idea of CA-RSS is to use simultaneously visual and strength information of visible lights to achieve accurate positioning at low computational cost. With rectangular LED layout, the irradiance angles of the visible lights can be estimated by visual information. Based on the estimated irradiance angles, the distances between the LEDs and the receiver can be obtained. Then, the two-dimensional (2D) position of the receiver can be calculated using a linear least square method. In addition, based on the solid geometry theory, we propose an independent solution method for three-dimensional (3D) positioning, which can avoid error propagation issue when compared with dependent solution method. To extend further CA-RSS for scenarios having different LED orientations and heights, we then propose a generalized CA-RSS algorithm for broader applications. Moreover, considering the imperfect channel model in practice, we propose a calibration method of Lambertian model for practical uses. Simulation results show that the proposed algorithm can achieve 90th percentile accuracies of less than 12.5 cm for 3D positioning regardless of the orientation of the receiver and the orientation and the height of the LEDs. In addition, experimental results show that using the calibration method, the proposed algorithm achieves an average 3D positioning error of 6.31 cm with different receiver orientations.

中文翻译：

---

多角度相机辅助可见光定位接收信号强度算法


提出了一种用于可见光定位（VLP）系统的相机辅助接收信号强度（CA-RSS）算法。 CA-RSS的基本思想是同时利用可见光的视觉和强度信息以较低的计算成本实现精确定位。通过矩形LED布局，可以通过视觉信息估计可见光的辐照角。根据估计的辐照角，可以获得 LED 和接收器之间的距离。然后，可以使用线性最小二乘法计算接收器的二维（2D）位置。此外，基于立体几何理论，我们提出了一种三维（3D）定位的独立求解方法，与依赖求解方法相比，可以避免误差传播问题。为了进一步将 CA-RSS 扩展到具有不同 LED 方向和高度的场景，我们提出了一种适用于更广泛应用的通用 CA-RSS 算法。此外，考虑到实际中信道模型的不完善，我们提出了一种实用的朗伯模型校准方法。仿真结果表明，无论接收器的方向以及 LED 的方向和高度如何，所提出的算法都可以实现 3D 定位的 90% 精度小于 12.5 cm。此外，实验结果表明，利用标定方法，该算法在不同接收方位下的平均3D定位误差为6.31 cm。