Bionic target localization method based on range-azimuth joint dictionary

The traditional underwater sonar system usually achieve high angle resolution by increasing array aperture and the number of array elements, but this method will inevitably lead to complex system and high cost. Given that big brown bats have obtained surprisingly high resolution using a simple system, this paper proposes a bionic target localization method. First, a range-azimuth joint dictionary was constructed based on the bionic system of multi-harmonic emission and double random array reception. Then, the coherence characteristic of the dictionary was analyzed and the range and azimuth of the target were estimated, and at last the experimental verification was completed. The results show that the bionic range-azimuth joint estimation based on sparse signal representation can achieve high-precision target localization under the condition of echo high aliasing.

Sensitivity Analysis of a Beamforming Technique for Acoustic Measurements

Beamforming is a technique that is used to determine the location of an acoustic source and the sound level spectrum of the signal produced by the source. This technique involves an array of microphones which record acoustic signals at multiple locations. A detailed analysis of the beamforming technique was carried out for three different array geometries: a uniform linear array, a uniform planar array, and a random array. The effect of various parameters, such as the number of microphones in an array, on the applicability of the technique was examined using both simulations and experiments. The simulation results established that the source localization capability of a uniform linear array is limited to an acoustic source lying in the plane of the array. In contrast, a planar array (either uniform or random) does not suffer the above limitation. The results also showed that a random array (eg., a spiral array) is the best of all the array geometries. The experimental results demonstrated the robustness of the beamforming technique in localizing an acoustic source and also confirmed the superiority of a uniform planar array over a uniform linear array.

Implementation of a novel reactive navigation algorithm

In this project a reactive navigation algorithm is applied to a non-holonomic differential drive robot. The algorithm uses a stochastic process to navigate a robot through terrain while lacking a priori information. A graph is made from a random array of points that is used to connect the current location of the robot to its destination. Dijkstra's algorithm is used to select the shortest route that leads to the destination. The robot attempts to traverse this route until it detects that it is being blocked by an obstacle. The graph is then recreated with different random points, an a new route is calculated. This procedure is repeated until the robot arrives at its destination. This is tested by making a simulated robot with perfect localization travel through two kinds of environments. Processing speed is maintained by hashing location information according to its coordinates.

Sensitivity Analysis of a Beamforming Technique for Acoustic Measurements

Beamforming is a technique that is used to determine the location of an acoustic source and the sound level spectrum of the signal produced by the source. This technique involves an array of microphones which record acoustic signals at multiple locations. A detailed analysis of the beamforming technique was carried out for three different array geometries: a uniform linear array, a uniform planar array, and a random array. The effect of various parameters, such as the number of microphones in an array, on the applicability of the technique was examined using both simulations and experiments. The simulation results established that the source localization capability of a uniform linear array is limited to an acoustic source lying in the plane of the array. In contrast, a planar array (either uniform or random) does not suffer the above limitation. The results also showed that a random array (eg., a spiral array) is the best of all the array geometries. The experimental results demonstrated the robustness of the beamforming technique in localizing an acoustic source and also confirmed the superiority of a uniform planar array over a uniform linear array.

Implementation of a novel reactive navigation algorithm

In this project a reactive navigation algorithm is applied to a non-holonomic differential drive robot. The algorithm uses a stochastic process to navigate a robot through terrain while lacking a priori information. A graph is made from a random array of points that is used to connect the current location of the robot to its destination. Dijkstra's algorithm is used to select the shortest route that leads to the destination. The robot attempts to traverse this route until it detects that it is being blocked by an obstacle. The graph is then recreated with different random points, an a new route is calculated. This procedure is repeated until the robot arrives at its destination. This is tested by making a simulated robot with perfect localization travel through two kinds of environments. Processing speed is maintained by hashing location information according to its coordinates.