| Polarized-light-aided navigation is a swift and convenient navigational approach inspired by the biological studies on the homing mechanisms of some hymenoptera such as sand ants, bees, locusts and crickets, etc. Falling into the category of visual navigation, this approach relies heavily on the skylight polarization pattern, one of the inherent attributes of the earth and featuring steady distribution, which makes it invulnerable to man-made sabotage and interference in short term and qualified for the autonomous navigation in alien environment with weak or even zero satellite signals. In light of these advantages, the approach has a brilliant prospect in military and civilian fields. Therefore, combining the modern navigational theories and technologies with the bionic research on insects’ utilization of the polarized light for orientation, we propose the autonomous navigational approach mainly by calculating the vehicle’s spatial attitude parameters with the skylight polarization pattern.The polarized-light-aided navigation derives from the studies of the biological mechanisms by which insects utilize polarized light for navigation. We discuss the relative literature about the behavioral, anatomic and electrophysiological studies on insects’ way of navigation and figure out the biological model of polarized light navigation. Through analyzing the biological mechanism of insects’ polarization-sensitive ommatidia in the dorsal rim area of their compound eyes, we make it possible the manufacturability of the four-channel polarization sensor, which lay the foundation for the further development of the polarization navigation.How to obtain the skylight polarization information and process it properly is the key to accurate navigation. By use of Matlab software, we duplicate the Rayleigh scattering model and work out the ideal skylight polarization pattern with the clear weather condition. Two set of detecting devices are established with a rotating platform and an on-board polarization analyzer, a panoramic lens and single lens reflex. We obtain the actual skylight polarization information through point by point scanning and image processing. Clustering method is introduced in this research to analyze the complicated polarization data. By this way, the issue of pursuing the astronomical markers is transformed into the determination of corresponding category center of polarization degree and polarization angle, and the optimum matching of the data is worked out through alternate optimization.Since the timely capture of the attitude parameters is the key to vehicle control during the flight, we focus our major concern on this. With the detected polarization data, we extract the information of solar projection, solar meridian, and zenith point and thus simulate the entire structure of polarization pattern. By use of the data of these astronomical markers and polarization pattern, we set up three mathematical models to determine the vehicle’s spatial attitude. By tracing the position variation of these markers with the carrier’s coordinate system, the attitude rotation matrix is established and the attitude parameters calculated.We test the three models theoretically with the detected polarization data. A virtual vehicle control panel is set up to simulate the attitude variation of vehicles when they rotate three-dimentionally in the air. For each simulation practice, we arbitrarily designate two of the three attitude angles and have the third one shifting from to. Finally, the relative attitude variation and root-mean-square error is calculated and analyzed with our method.In sum, the bio-inspired polarized-light-aided navigational approach proposed in this contribution is not simply a duplication of the biological model. It is an original approach, combining the advanced navigational devices with the biological utilization. Therefore, it can be utilized to be not only the auxiliary device of conventional approach, but also an independent navigational control for vehicles. |
PDF Full Text Request