Polarized skylight compass based on a soft-margin support vector machine working in cloudy conditions

Ultraviolet bionic compass method based on non-ideality correction and statistical guidance in twilight conditions

Bionic polarization compass is a fascinating subject in the navigation domain. However, the polarization navigation accuracy is severely degraded by the influence of city glow at dusk. Therefore, we proposed an ultraviolet bionic compass method based on non-ideality correction and statistical guidance. A non-ideal polarization imaging model was established to correct the system detection error. A meridian extraction algorithm based on the statistical properties of solar direction vectors was proposed for accurate heading calculation. The proposed algorithm was demonstrated experimentally and reduced the heading error to approximately 1°, which shows strong anti-interference performance against urban glare.

Passive Polarized Vision for Autonomous Vehicles: A Review

This review article aims to address common research questions in passive polarized vision for robotics. What kind of polarization sensing can we embed into robots? Can we find our geolocation and true north heading by detecting light scattering from the sky as animals do? How should polarization images be related to the physical properties of reflecting surfaces in the context of scene understanding? This review article is divided into three main sections to address these questions, as well as to assist roboticists in identifying future directions in passive polarized vision for robotics. After an introduction, three key interconnected areas will be covered in the following sections: embedded polarization imaging; polarized vision for robotics navigation; and polarized vision for scene understanding. We will then discuss how polarized vision, a type of vision commonly used in the animal kingdom, should be implemented in robotics; this type of vision has not yet been exploited in robotics service. Passive polarized vision could be a supplemental perceptive modality of localization techniques to complement and reinforce more conventional ones.

Image-registration-based solar meridian detection for accurate and robust polarization navigation

Skylight polarization, inspired by the foraging behavior of insects, has been widely used for navigation for various platforms, such as robots, unmanned aerial vehicles, and others, owing to its stability and non-error-accumulation. Among the characteristics of skylight-polarized patterns, the angle of polarization (AOP) and the degree of polarization (DOP) are two of the most significant characteristics that provide abundant information regarding the position of the sun. In this study, we propose an accurate method for detecting the solar meridian for real-time bioinspired navigation through image registration. This method uses the AOP pattern to detect the solar meridian and eliminates the ambiguity between anti-solar meridian and solar meridian using the DOP pattern, resulting in an accurate heading of the observer. Simulation experiments demonstrated the superior performance of the proposed method compared to the alternative approaches. Field experiments demonstrate that the proposed method achieves real-time, robust, and accurate performance under different weather conditions with a root mean square error of 0.1° under a clear sky, 0.18° under an overcast sky with a thin layer of clouds, and 0.32° under an isolated thick cloud cover. Our findings suggest that the proposed method can be potentially used in skylight polarization for real-time and accurate navigation in GPS-denied environments.

Measurement Modeling and Performance Analysis of a Bionic Polarimetric Imaging Navigation Sensor Using Rayleigh Scattering to Generate Scattered Sunlight

The bionic polarimetric imaging navigation sensor (BPINS) is a navigation sensor that provides absolute heading, and it is of practical engineering significance to model the measurement error of BPINS. The existing BPINSs are still modeled using photodiode-based measurements rather than imaging measurements and are not modeled systematically enough. This paper proposes a measurement performance analysis method of BPINS that takes into account the geometric and polarization errors of the optical system. Firstly, the key error factors affecting the overall measurement performance of BPINS are investigated, and the Stokes vector-based measurement error model of BPINS is introduced. Secondly, based on its measurement error model, the effect of the error source on the measurement performance of BPINS is quantitatively analyzed using Rayleigh scattering to generate scattered sunlight as a known incident light source. The numerical results show that in angle of E-vector (AoE) measurement, the coordinate deviation of the principal point has a greater impact, followed by grayscale response inconsistency of CMOS and integration angle error of micro-polarization array, and finally lens attenuation; in degree of linear polarization (DoLP) measurement, the grayscale response inconsistency of CMOS has a more significant impact. This finding can accurately guide the subsequent calibration of BPINS, and the quantitative results provide an important theoretical reference for its optimal design.

Skylight Polarization Pattern Simulator Based on a Virtual-Real-Fusion Framework for Urban Bionic Polarization Navigation

In a data-driven context, bionic polarization navigation requires a mass of skylight polarization pattern data with diversity, complete ground truth, and scene information. However, acquiring such data in urban environments, where bionic polarization navigation is widely utilized, remains challenging. In this paper, we proposed a virtual-real-fusion framework of the skylight polarization pattern simulator and provided a data preparation method complementing the existing pure simulation or measurement method. The framework consists of a virtual part simulating the ground truth of skylight polarization pattern, a real part measuring scene information, and a fusion part fusing information of the first two parts according to the imaging projection relationship. To illustrate the framework, we constructed a simulator instance adapted to the urban environment and clear weather and verified it in 174 urban scenes. The results showed that the simulator can provide a mass of diverse urban skylight polarization pattern data with scene information and complete ground truth based on a few practical measurements. Moreover, we released a dataset based on the results and opened our code to facilitate researchers preparing and adapting their datasets to their research targets.

In-Flight Alignment of Integrated SINS/GPS/Polarization/Geomagnetic Navigation System Based on Federal UKF

As a common integrated navigation system, the strapdown inertial navigation system (SINS)/global positioning system (GPS) can estimate velocity and position errors well. Many auxiliary attitude measurement systems can be used to improve the accuracy of attitude angle errors. In this paper, the in-flight alignment problem of the integrated SINS/GPS/Polarization/Geomagnetic navigation system is discussed. Firstly, the SINS/Geomagnetic subsystem is constructed to improve the estimation accuracy of horizontal attitude angles. Secondly, the polarization sensor is used to improve the estimation accuracy of heading angle. Then, a federal unscented Kalman filter (FUKF) with non-reset structure is applied to fuse the navigation data. Finally, simulation results for the integrated navigation system are provided based on experimental data. It can be shown that the proposed approach can improve not only the speed and position, but also the attitude error effectively.

Bionic orientation method based on polarization imaging in HDR scenes

An increasing number of bio-inspired navigation approaches have been designed based on polarization cameras. However, digital cameras can sense a much narrower field of vision than the vision of insects or human beings. In this study, we propose an adaptive skylight polarized orientation method for high dynamic range (HDR) scenes. Initially, we built a model of the image acquisition pipeline that can recover HDR irradiance maps from polarization images. Subsequently, the orientation method was designed based on a combination of the irradiance maps and the least squares methods. Some preprocessing steps were utilized to eliminate occlusion interference. In addition, an autoexposure adjustment method was proposed using information entropy and heuristic segmentation. Finally, the experimental results show that the proposed method can improve the accuracy of bionic orientation and adaption to skylight with occlusions and interference in natural conditions.

Three-dimensional attitude determination strategy for fused polarized light and geomagnetism

Using polarized light sensors to obtain only two-dimensional heading information does meet actual needs in navigation. Instead, an alternative method is proposed that uses the positional information of the Sun and geomagnetic information to calculate the three-dimensional attitude of a vehicle. First, the theoretical background of the polarization mode of skylight is described, and the scheme in using the atmospheric polarization pattern to calculate the solar position is presented. Second, the traditional three-axis attitude-determination (TRIAD) algorithm that exploits the solar position vector and the geomagnetic vector to obtain the three-dimensional attitude and the optimized TRIAD algorithm are introduced. Static and turntable experiments are described that verify the accuracy of the attitude calculation. Experimental results show that when using the optimized TRIAD algorithm, the root mean square errors for the roll angle, pitch angle, and heading angle are 0.1225°, 0.668°, and 1.0234°, respectively. This means that the optimized TRIAD algorithm performs significantly better than the traditional TRIAD algorithm and demonstrates that using the solar position and the geomagnetic information to obtain the three-dimensional attitude of the vehicle is very effective.

Measurement error model of the bio-inspired polarization imaging orientation sensor

This article studies the measurement error model and calibration method of the bio-inspired polarization imaging orientation sensor (BPIOS), which has important engineering significance for promoting bio-inspired polarization navigation. Firstly, we systematically analyzed the measurement errors in the imaging process of polarized skylight and accurately established an error model of BPIOS based on Stokes vector. Secondly, using the simulated Rayleigh skylight as the incident surface light source, the influence of multi-source factors on the measurement accuracy of BPIOS is quantitatively given for the first time. These simulation results can guide the later calibration of BPIOS. We then proposed a calibration method of BPIOS based on geometric parameters and the Mueller matrix of the optical system and conducted an indoor calibration experiment. Experimental results show that the measurement accuracy of the calibrated BPIOS can reach 0.136°. Finally, the outdoor performance of BPIOS is studied. Outdoor dynamic performance test and field compensation were performed. Outdoor results show that the heading accuracy of BPIOS is 0.667°.

Adaptive method for estimating information from a polarized skylight

The acquisition and processing of skylight polarization information forms the cornerstone in modern navigation systems that are developed by imitating certain biological mechanisms. The accuracy of skylight polarization mode information plays a major part in improving the accuracy of polarized light navigation. This paper mainly focuses on developing a methodology that can avoid the error caused by the inaccurate rotation of the polarizer and manual readings from non-electrical equipment, when the time-sequence polarization measurement system is used to obtain the skylight polarization mode information. We propose an adaptive algorithm that can obtain the pictures of angle of polarization and degree of polarization with sets of random rotation angles with no need for precise readings for the rotation angle of the polarizer. By allocating initial random values to rotation angles, a simple iterative estimation method like the Gaussian-Newton method can be used to converge calculated angle of polarization and degree of polarization values to their respective real values. The experiment results show that the proposed method can be used to estimate polarization information with high accuracy and universality under various experiment settings including both sunny and cloudy weathers. Meanwhile, the time efficiency of the proposed method is comparable to traditional methods.

Design and calibration model of a bioinspired attitude and heading reference system based on compound eye polarization compass

Insects such as honeybees are capable of fusing the information sensed by multiple sensory organs for attitude and heading determination. In this paper, inspired by the sensory fusion mechanism of insects' polarization compass and haltere, a bioinspired polarization-based attitude and heading reference system (PAHRS) is presented. The PAHRS consists of compound eye polarization compass and inertial measurement unit (IMU). By simulating multi-view structure of the dorsal rim area in insects' compound eyes, a non-coplanar 'polarization-opponent (POL)-type' architecture is adopted for the compound eye polarization compass. The polarization compass has multi-directional observation channels, which is capable of adaptively selecting the angle of polarization and obtaining the polarization vectors. Therefore, the environmental adaptability of the polarization compass can be enhanced. In addition, the integration strategy between the compound eye polarization compass and IMU is proposed. Moreover, the sources of system errors are analyzed to improve the heading angle accuracy, based on which a new calibration model is established to compensate the installation errors of the PAHRS. Finally, experiments are carried out under both clear sky and cloudy conditions. The test results show that the error root mean square of heading angle is 0.14° in clear sky, and 0.42° in partly cloudy conditions.