9 January 2020 in IEEE Transactions on Biomedical Circuits and Systems
In this work, a memristive circuit with affective multi-associative learning function is proposed, which mimics the process of human affective formation. It mainly contains three modules: affective associative learning, affective formation, affective expression. The first module is composed of several affective single-associative learning circuits consisting of memristive neurons and synapses. Memristive neuron will be activated and output pulses if its input exceeds the threshold. After it is activated, memristive neuron can automatically return to the inactive state. Memristive synapse can realize learning and forgetting functions based on the signals from pre- and post-neurons. The learning rule is pre-neuron activated lags behind post-neuron for a short time; the forgetting rule is to repeatedly activate pre-neuron after the emotion is learned. The process of learning or forgetting corresponds to facilitating or inhibiting synaptic weight, that is, decreasing or increasing memristance continuously. Different voltage signals applied to memristors and different parameters of memristors would lead to different synaptic weights which indicate different affective association. The second module can convert affective signals to corresponding emotions. The formed emotions can be shown in a face by the third module. The simulation results in PSPICE show that the proposed circuit system can learn, forget and form emotions like human. If the proposed circuit is further used on a humanoid robot platform through further research, the robot will have the ability of affective interaction with human so that it can be effectively used in affective company and other aspects.
1 January 2018 in Complexity
Adaptive Monte Carlo localization (AMCL) algorithm has a limited pose accuracy because of the nonconvexity of the laser sensor model, the complex and unstructured features of the working environment, the randomness of particle sampling, and the final pose selection problem. In this paper, an improved AMCL algorithm is proposed, aiming to build a laser radar‐based robot localization system in a complex and unstructured environment, with a LIDAR point cloud scan‐matching process after the particle score calculating process. The weighted mean pose of AMCL particle swarm is used as the initial pose of the scan matching process. The LIDAR point cloud is matched with the probability grid map from coarse to fine using the Gaussian‐Newton method, which results in more accurate poses. Moreover, the scan‐matching pose is added into the particle swarm as a high‐weight particle. So the particle swarm after resampling will be more concentrated in the correct position. The particle filter and the scan‐matching process form a closed loop, thus enhancing the localization accuracy of mobile robots. The experiment results demonstrate that the proposed improved AMCL algorithm is superior to the traditional AMCL algorithm in the complex and unstructured environment, by exploiting the high‐accuracy characteristic of scan matching while inheriting the stability of AMCL.
27 July 2020 in IEEE Sensors Journal
Considering the adverse impact of speed measurement on the accuracy of pose estimation after a mobile robot slips, collides, or abducts, this paper proposes a monocular inertial simultaneous localization and mapping algorithm that includes wheel speed anomaly detection. The algorithm adds wheel speed measurement to the least squares problem in a tightly coupled manner and uses a nonlinear optimization method to maximize the posterior probability to solve the optimal state estimation. For the speed control of the Mecanum wheel, because the existing closed-loop speed control method cannot calculate the motion constraint error, this paper reports a design of a control method of the Mecanum wheel moving chassis based on torque control, which can use the motion constraint error to estimate the credibility of the wheel speed measurement to detect whether the chassis movement status is abnormal; meanwhile, to prevent the chassis speed measurement error from adversely affecting the robot pose estimation, this paper uses three methods to actively detect whether the chassis movement is abnormal, and analyze the chassis movement status in real time. When it is determined that the chassis has abnormal motion, the wheel odometer pre-integration measurement of the current frame is actively removed from the state estimation equation, thereby ensuring the accuracy of the pose estimation. Experimental results show the feasibility and effectiveness of the method proposed in this paper, and the algorithm is robust.