Search

With recent breakthroughs in the electrode-based brain-computer-interface technology, cognition enhancement becomes increasingly feasible by applying external stimuli to neurons in the brain. An essential stepstone for this objective is to develop compatible representations between artificial and biological neural networks. We propose that grid cells can serve as such an interfacing representation. First, we empirically proved that DNNs replicated the grid cell behaviors on non-spatial cognitive tasks. Second, we developed theoretical analysis on the formation of hexagonal activation pattern of grid cells, which is independent of physical implementations of neural networks. Third, we performed extensive experiments to validate our proposal.

Localization for moving multi-target in urban environment is difficult because of the ambiguities of targets’ different multipaths. This paper considers the targets localization problem via non-line-of-sight (NLOS) signals in an L-shaped corner environment using the time-division multiple-input-multiple-output (TD-MIMO) imaging radar. We first analyse the propagation characters of the multipath signals and establish the signal model of targets hiding behind corner for TD-MIMO. Then, a novel multi-target localization algorithm by exploiting the multipath ghosts after applying back-projection (BP) imaging algorithm is proposed. Finally, we illustrate the validation the proposed algorithm via several simulation results.

This paper addresses a target imaging method for non-line-of-sight (NLOS) target shielded by surrounding in the L-junction scenario, exploiting ultra wideband (UWB) multiple-input multiple-output (MIMO) radar. We analyze the electromagnetic wave diffraction and reflection propagation in the L-junction scenario and establish the echo model. An NLOS target imaging method based on multipath exploitation of the diffraction and first-order reflection is proposed. Moreover, we also present a multiplication fusion method to suppress the high-order multipath ghosts. Finally, simulation results validate the validity of the multipath-exploitation-based target imaging method.

In this paper, we consider a corner target localization problem using a monostatic L-band ultra-wideband (UWB) radar with one transmitting antenna and one receiving antenna. We first analyse the characters of the measured signal received by an L-band UWB radar. Then, we propose a target localization algorithm using times of arrival (TOAs) of multipath components produced by diffraction and specular reflections on surrounding surfaces. The core part of this algorithm is an association method which can correctly match the measured TOAs with the propagation paths. This method can cope with the case when some missed detections of TOAs appear. Simulation and experimental results validate the proposed algorithm.

This paper addresses a localization problem for a single non-line-of-sight (NLOS) target shielded by a rectangular building corner in an urban environment exploiting an ultra-wideband (UWB) radar. A novel electromagnetic waves propagation model referring to such building corner is established. Specifically, we firstly analyze the characteristics of the radar echoes propagating along the direct diffraction path and the side wall reflection path. Then, we extract the lengths of these two paths after the target detecting via CFAR technique. Next, based on the priori-known building layout, we obtain the position of the hidden target accurately. Finally, numerical simulations and experimental data validate the proposed approach.

In this paper the power requirements of deception jamming in the signal acquisition phase of BeiDou navigation receiver are discussed in two cases. In the first case, the deception jamming signal is released before the start of the receiver, with the method of analyzing the signal acquisition probability model and the acquisition correlation peak, the conclusion is obtained that the effective power gain of deception jamming signal is not less than 10dB. In the other case, the deception jamming signal is released after the start of the receiver, through analysis of the change of signal to noise ratio, the conclusion is obtained that the effective power gain of deception jamming signal is not less than 12dB. And the validity of the two conclusions is verified by experiments. The research of this paper has the important application value and military significance for the deception jamming detection and defense of BeiDou navigation terminal.

This paper presents a test bed for a wideband alldigital array radar, which consists of 32 linear array apertures. The operating frequency band is 1–3GHz and the instantaneous bandwidth is 500MHz. High-speed direct RF digitization and powerful digital signal processors are utilized in the design of the system. The test bed has been implemented now with a 32-elements array antenna and 8 digital receiver channels. Some experiments have been conducted and the testing results are presented in this paper. The testing results confirm the feasibility of the test bed, and guide us to carry on the remaining work of the whole system designing.

This paper presents an innovative locomotion concept with virtual center of motion (VCM) which can improve the performance of the mobile robot on rough terrain. The mechanism with the character of VCM is one of the limited degree of freedom (DOF) mechanisms in nature. This allows for simplifying the static model of the mobile robot. The VCM was applied to mobile robots to extend their climbing capabilities on terrain for its features. The front jockey mechanism of the robot was optimized through analyzing the relationship between the VCM and the level of the front wheel axis. In order to minimize the friction coefficient of the wheels, the parameters and the suspension position of the spring on the jockey were optimized through static analysis. An optimized prototype has been designed and developed. Experimental results show that the robot is able to passively overcome obstacles approximately up to 2.1 times its wheel diameter and climb up steps over 1.27 times its wheel diameter.

This paper presents an innovative locomotion concept for rough terrain based on six motorized rimless wheels. Different from traditional wheels, the rimless wheels can not only provide more traction on uneven earth's surface, but also have advanced adaptation to the complex terrain. In order to make the robot move stably and efficiently, the design proposes a novel mechanism of the rimless-wheeled robot and motion planning, including running and turning. Especially, the control strategy takes the motion planning of an unequal number of spokes on each wheel into consideration. In the experiments, the results show that the robot is able to climb about 1~1.5 times high stairs and travel on rough terrain without its center of gravity up and down intensely. Furthermore, some motion parameters are optimized to improve the efficiency of the robot during its running and turning.