Chongqing University

Star Institute of Intelligent Systems

Dibi (chongqing) Intelligent Technology Research Institute Co., Ltd.

University of Electronic Science and Technology of China

Db (chongqing) Intelligent Technology Research Institute Co., Ltd.

Chongqing Yingdi Industry (group) Co., Ltd.

Chongqing University, Star Institute of Intelligent Systems, Db (chongqing) Intelligent Technology Research Institute Co., Ltd.

Star (chongqing) Intelligent Equipment Technology Research Institute Co., Ltd.

Yongduan Song

Junfeng Lai

Ziqiang Jiang

Li Huang

Hong Long

Shilei Tan

Huan Liu

Jie Zhang

Fang Hu

Rui Li

Jiangyu Wu

Lihui Tan

Feng Yang

Huan Chen

Jian Zhang

Chunxu Ren

Yating Yang

Li Chen

Zhengtao Yu

Siyu Li

Qin Hu

Shichun Wang

Yiwen Zhang

Jiawei Chen

Zhixi Shen

Shengtao Pan

Congyi Zhang

Jian Liu

Jie Lei

Kuilong Yang

Ge Song

Bowen Zhang

Wenqi Li

Xin Zhou

Shuaicheng Hou

Haoyuan Zhong

Hongyu Xia

Xiao Chen

Mi Fang

Qin Chen

Shihao Pu

Ke'er Chen

Qingchao Song

Caishi He

Haifeng Liu

Yujuan Wang

Mingyang Liu

Saiyu Wang

Shuai Wang

Gonglin Lu

Zhilin Wang

Lei Fang

Yankai Zhang

Xiao Cao

Xiaohu Pan

Han Bao

Yingrui Chen

The present disclosure discloses an adaptive high-precision compression method and system based on a convolutional neural network model, and belongs to the fields of artificial intelligence, computer vision, and image processing. According to the method of the present disclosure, coarse-grained pruning is performed on a neural network model by using a differential evolution algorithm first, and the coarse-grained space is quickly searched through an entropy importance criterion and an objective function with good guidance to obtain a near-optimal neural network structure. Then fine-grained search space is built on the basis of an optimal individual obtained from the coarse-grained search, and fine-grained pruning is performed on the neural network model by a differential evolution algorithm to obtain a network model with an optimal structure. Finally, the performance of the optimal model is restored by using a multi-teacher multi-step knowledge distillation network to reach the precision of an original model.

Adaptive high-precision compression method and system based on convolutional neural network model

Disclosed is a snow shoveling and snow discharging assembly of a snow sweeping robot. The snow shoveling and snow discharging assembly comprises a snow stirring structure, a snow feeding structure and a snow raising structure. Stirring cutters in the snow stirring mechanism are driven by a stirring cutter shaft to stir bottom area snow into a snow feeding pipe of the snow feeding structure, an air blower in the snow feeding structure blows accumulated snow into a snow raising pipe of the snow raising structure through a connecting pipe, the snow raising pipe can be driven by a steering motor to rotate by a certain angle to control the snow raising direction, and a second electric push rod acts to control pitching of the guide part, so that the height of snow during snow raising is controlled.

Snow shoveling and snow discharging assembly of snow sweeping robot

The invention discloses an adaptive path planning method based on neutral networks trained by the evolutional algorithms, the neutral network training method comprises input and output of the data acquired by the mobile sensors installed on the mobile robots as the neutral networks, and training and optimization of the recurrent neutral networks based on the evolutional algorithms; the path planning method refers to the application of the trained neutral networks to the path planning of the mobile robot, the invention effectively improves local quick search capability and global search capability of the algorithms by applying the evolutional algorithms to the optimization of the recurrent neutral networks, so that the robot can plan a rational path in a dense and uncertain environment.

Adaptive path planning method based on neutral networks trained by the evolutional algorithms

The present disclosure provides a decentralized active equalization method for a cascaded lithium-ion battery pack. The method includes: connecting each battery cell in the cascaded lithium-ion battery pack to a direct current (DC) bus through an equalizer respectively, where each equalizer includes an independent controller, a sampling circuit, a power supply circuit, a drive circuit, and a main circuit; connecting an input terminal of the main circuit to a corresponding battery cell, and connecting an output terminal of the main circuit to the DC bus. The present disclosure solves the technical problem that an existing cascaded lithium-ion battery pack equalization method cannot achieve equalization when a centralized controller failure or a communication failure occurs, improves the reliability of the equalization method, can make the equalizer work at high efficiency by configuring parameters of C, K, and R, and speeds up the equalization or improves the equalization accuracy. The present disclosure uses a bus structure, which can achieve the equalization of all battery cells at the same time, thereby greatly improving the equalization speed. The present disclosure uses a modular design and can be arbitrarily configured according to the scale of the cascaded lithium battery pack, and has strong scalability.

Decentralized active equalization method for cascaded lithium-ion battery pack

Embodiments of the present disclosure provide a quantitative method and system for attention based on a line-of-sight estimation neural network, which improves the stability and training efficiency of the line-of-sight estimation neural network. A few-sample learning method is applied to training of the line-of-sight estimation neural network, which improves generalization performance of the line-of-sight estimation neural network. A nonlinear division method for small intervals of angles of the line of sight is provided, which reduces an estimation error of the line-of-sight estimation neural network. Eye opening and closing detection is added to avoid the line-of-sight estimation error caused by an eye closing state. A method for solving a landing point of the line of sight is provided, which has high environmental adaptability and can be quickly used in actual deployment.

Quantitative analysis method and system for attention based on line-of-sight estimation neural network

The present disclosure provides a real-time vehicle overload detection method based on a convolutional neural network (CNN). The present disclosure detects a road driving vehicle in real time with a CNN method and a you only look once (YOLO)-V3 detection algorithm, detects the number of wheels to obtain the number of axles, detects a relative wheelbase, compares the number of axles and the relative wheelbase with a national vehicle load standard to obtain a maximum load of the vehicle, and compares the maximum load with an actual load measured by a piezoelectric sensor under the vehicle, thereby implementing real-time vehicle overload detection. The present disclosure has desirable real-time detection, can implement no-parking vehicle overload detection on the road, and avoids potential traffic congestions and road traffic accidents.

Real-time vehicle overload detection method based on convolutional neural network

The present disclosure provides a neural network-based method for calibration and localization of an indoor inspection robot. The method includes the following steps: presetting positions for N label signal sources capable of transmitting radio frequency (RF) signals; computing an actual path of the robot according to numbers of signal labels received at different moments; computing positional information moved by the robot at a tmoment, and computing a predicted path at the tmoment according to the positional information; establishing an odometry error model with the neural network and training the odometry error model; and inputting the predicted path at the tmoment to a well-trained odometry error model to obtain an optimized predicted path. The present disclosure maximizes the localization accuracy for the indoor robot by minimizing the error of the odometer with the odometry calibration method.

Neural network-based method for calibration and localization of indoor inspection robot

The present disclosure relates to a method for recognizing facial expressions based on adversarial elimination. First, a facial expression recognition network is built based on a deep convolutional neural network. On a natural facial expression data set, the facial expression recognition network is trained through a loss function to make facial expression features easier to distinguish. Then some key features of input images are actively eliminated by using an improved confrontation elimination method to generate a new data set to train new networks with different weight distributions and feature extraction capabilities, forcing the network to perform expression classification discrimination based on more features, which reduces the influence of interference factors such as occlusion on the network recognition accuracy rate, and improving the robustness of the facial expression recognition network. Finally, the final expression classification predicted results are obtained by using network integration and a relative majority voting method.

Method for recognizing facial expressions based on adversarial elimination

The present disclosure provides a method for object detection and recognition based on a neural network. The method includes: adding a detection layer following three detection layers of an existing YOLOv5 network model, to construct a new YOLOv5 network model; then, training the new YOLOv5 network model by considering an overlapping area between a predicted box and a ground truth box, a center-to-center distance between the two boxes, and an aspect ratio of the two boxes; and finally, inputting a to-be-detected image into the trained new YOLOv5 network model, outputting a predicted box of an object and probability values corresponding to a class to which the object belongs, and setting a class corresponding to a maximum probability value as a predicted class of the object in the to-be-detected image. This method can quickly and effectively detect multiple classes of objects. Especially, a detection effect for small objects is more ideal.

Method for object detection and recognition based on neural network

A method for obstacle detection and recognition for an intelligent snow sweeping robot is disclosed, comprising: 1) disposing ultrasonic sensors at a front end of the snow sweeping robot to detect distance information from an obstacle ahead; and disposing radar sensors at the front and rear of the snow sweeping robot to detect whether a creature suddenly approaches; 2) processing signals detected by each of the ultrasonic sensors and radar sensors, and calculating a forward distance of the snow sweeping robot; and 3) determining a snow cover extent of a working road, detecting a change of the distance from the obstacles, and recognizing the obstacles for conditions of an ultrasonic ranging variation ratio and a variation of the forward distance of the snow sweeping robot, a change of the signal detected by radar sensors, and a descriptive statistic of the snow cover extent within a specific time period.

Method for obstacle detection and recognition for intelligent snow sweeping robot

The present disclosure discloses a neural network adaptive tracking control method for joint robots, which proposes two schemes: robust adaptive control and neural adaptive control, comprising the following steps: 1) establishing a joint robot system model; 2) establishing a state space expression and an error definition when taking into consideration both the drive failure and actuator saturation of the joint robot system; 3) designing a PID controller and updating algorithms of the joint robot system; and 4) using the designed PID controller and updating algorithms to realize the control of the trajectory motion of the joint robot. The present disclosure may solve the following technical problems at the same time: the drive saturation and coupling effect in the joint system, processing parameter uncertainty and non-parametric uncertainty, execution failure handling during the system operation, compensation for non-vanishing interference, and the like.

Neural network adaptive tracking control method for joint robots

The present disclosure discloses a method for route optimization based on dynamic window and redundant node filtering, comprising using an existing raster map data set to determine the coordinate information of a starting position and a destination position of movement, and to mark a destination node and an obstacle node in the raster map; using A* algorithm to plan a global route; globally optimizing the global route planned by A* algorithm, and filtering redundant nodes out; combining a dynamic window algorithm to perform the local optimization section by section on the optimized global route so as to obtain a final global route. According to the present disclosure, the combination of algorithms reduces a single movement duration of a mobile robot and improves the smoothness of the movement route curve. At the same time, the problems of the robot occurring on the route during the static driving are alleviated.

Method for route optimization based on dynamic window and redundant node filtering

A snow shovel structure of a snow plow robot. The snow shovel structure includes a housing where a snow shovel mechanism is. The snow shovel mechanism extends outside the housing and includes a first motor fixed on a top of the housing. The first motor is fixedly connected with a telescopic rod through an output shaft. A second motor is further provided on a top portion of an inner chamber of the housing, and a horizontal plate is fixedly arranged on a side wall of the inner chamber of the housing.

Snow shovel structure of snow plow robot

The present disclosure discloses a method for quick weighing of loose and small packages of traditional Chinese medicine (TCM), comprising: 1) establishing an equivalent physical model of a weighing system for loose and small packages, and, through Laplace transformation and Z transformation, obtaining a formula of the mass M of medicine packages to be weighed; 2) calculating a, a, band bon the basis of a vector prediction error; 3) according to a, a, b, and bobtained and the formula of the mass of medicine packages to be weighed, obtaining the mass of the loose and small packages of TCM that are weighed. According to the method, the scalar prediction error is expanded into the vector prediction error based on the traditional method so as to construct a new identification model based on the vector prediction error, and as a result the utilization efficiency of prediction error turns higher.

Method for quick weighing of loose and small packages of traditional Chinese medicine (TCM)

A detection circuit for open, close and suspension states of a high and low level effective switch in a vehicle. The circuit includes an optocoupler circuit module, a low-level active path module, a high-level active path module, a filtering and debouncing module, a transient suppression module, and a wiring terminal. The optocoupler circuit module is connected to the low-level active path module, the high-level active path module and the low-level active path module are connected in parallel to the filtering and debouncing module, and the filtering and debouncing module is connected to the transient suppression module, and then connected to the external high-level active switch or low-level active switch through the wiring terminal. Whether it is a high-level active switch or a low-level active switch, the detection circuit can distinguish whether the switch is in the closed or suspended state, and the strong and weak voltages are isolated.

Detection circuit for open, close and suspension states of high and low level effective switch in vehicle

Growing community of inventors

Chongqing, China

Yongduan Song