Patrol Robot Research Articles

RGB camera-based fallen person detection system embedded on a mobile platform

Most injuries in the elderly are due to falls. The response time, to attend to the critically injured in such fall cases, is crucial to their survival. This paper presents a low-cost, autonomous assistive patrol robot which additionally includes a fallen person detection module with facial recognition that allows identification of patients. Patrol robots could be beneficial for care centers, where there is a considerable number of patients that require care. In these conditions, falls can be generally detected by the robotic platform during the post-fall phase. This allows the system to work with no frame rate constraints, allowing other tasks to be run simultaneously. Based on the YOLO network, we propose two approaches for the fallen person detector. The first approach can differentiate between fallen persons and persons doing ordinary activity in a single stage, while the second is a two-staged approach. The network weights were obtained using a fine-tuning process by retraining with our own extended Fall Person Dataset (E-FPDS), which we release as a benchmark for other RGB vision-based approaches. Quantitative evaluations confirm that the detector performs robustly in detecting fallen persons in different situations. The results also show a recall of 98.97% in our test set.

Improved Path Planning for Indoor Patrol Robot Based on Deep Reinforcement Learning

To solve the problems of poor exploration ability and convergence speed of traditional deep reinforcement learning in the navigation task of the patrol robot under indoor specified routes, an improved deep reinforcement learning algorithm based on Pan/Tilt/Zoom(PTZ) image information was proposed in this paper. The obtained symmetric image information and target position information are taken as the input of the network, the speed of the robot is taken as the output of the next action, and the circular route with boundary is taken as the test. The improved reward and punishment function is designed to improve the convergence speed of the algorithm and optimize the path so that the robot can plan a safer path while avoiding obstacles first. Compared with Deep Q Network(DQN) algorithm, the convergence speed after improvement is shortened by about 40%, and the loss function is more stable.

Autonomous Patrol Robot using Solid-State LiDAR with Narrow FoV

Autonomous Patrol Robot using Solid-State LiDAR with Narrow FoV

Vision Based Indoor Surveillance Patrol Robot Using Extended Dijkstra Algorithm in Path Plannin

Vision based patrol robot has been with great interest nowadays due to its consistency, cost effectiveness and no temperament issue. In recent times, Global positioning system (GPS) has been cooperated with Global Navigation Satellite System (GNSS) to come out with better accuracy quality in positioning, navigation, and timing (PNT) services to locate a device. However, such localization service is yet to reach any indoor facility. For an indoor surveillance vision based patrol robot, such limitation hinders its path planning capabilities that allows the patrol robot to seek for the optimum path to reach the appointed destination and return back to its home position. In this paper, a vision based indoor surveillance patrol robot using sensory manipulation technique is presented and an extended Dijkstra algorithm is proposed for the patrol robot path planning. The design of the patrol robot adopted visual type sensor, range sensors and Inertia Measurement Unit (IMU) system to impulsively update the map’s data in line with the patrol robot’s current path and utilize the path planning features to carry out obstacle avoidance and re-routing process in accordance to the obstacle’s type met by the patrol robot. The result conveyed by such approach certainly managed to complete multiple cycles of testing with positive result.

Patrol robot path planning in nuclear power plant using an interval multi-objective particle swarm optimization algorithm

This paper presents an interval multi-objective path planning (PP) scheme for patrol robot in nuclear power plant. The purpose of this PP scheme is to find collision-free paths with the shortest length and smallest risk degree. Firstly, a novel workspace modeling method is proposed to describe the static PP environment of patrol robot in nuclear power plant. Then considering the conflicts of the shortest length and smallest risk degree, an interval multi-objective particle swarm optimization (IMOPSO) method is used. In the IMOPSO, an ingenious interval update law for the particle’s global best position and local best position based on the crowding distance of each risk degree interval is used to increase the diversity of population, and an iterative procedure is adopted to update the particle’s position when the found paths are collided with some existing obstacles. Finally, three representative simulation tests are used to verify the validity of proposed IMOPSO method. Results show that comparing with other three well-known multi-objective evolutionary algorithms, our proposed method has the advantages of finding a better Pareto optimal paths.

Combined inspection strategy of bionic substation inspection robot based on improved Biological Inspired Neural Network

Aiming at the problem of complex and low efficiency of intelligent patrol paths in 220/500 kV bionic substations, this paper proposes a method for collaborative inspection of multiple patrol robots. Firstly, the deficiencies of the Biological Inspired Neural Network algorithm(BINN) in complex paths and large corner scenarios are analyzed, and an improved method for solving the neuron activity value near the boundary and an improved method for the neuron activity value near the obstacle are proposed, in order to solve the phenomenon that the decrease of neurons that can transmit activity values in the neighborhood leads to generally low neuron activity values, which leads to unreasonable path planning; Secondly, establish a substation area segmentation model based on the variable tangent method, divide the substation area into several sub-intervals without obstacles, and use the priority heuristic algorithm to select the robot travel path, so as to realize the multi-robot joint inspection task ; Finally, the method proposed in this paper reduces the path length, number of turns and rotation angle by 11.38%, 38,89% and 20.51, respectively, compared with the BINN algorithm, and by 5.75%, 18.18%, and 16.68 compared with the A* algorithm, which verifies the feasibility and effectiveness of the proposed method.

Graph-Based Knowledge Acquisition With Convolutional Networks for Distribution Network Patrol Robots

With the popularization of smart grids, patrol robots have become critical devices in the distribution networks to check the state of equipment. In order to enrich the knowledge of patrol robots in this complex scenario, this article presents the graph-based knowledge acquisition method with convolutional networks for distribution network patrol robots. The proposed method uses a graph convolutional network-based path-related embedding algorithm to complete the knowledge of the distribution network knowledge graph. The proposed algorithm generates the embeddings of entities and relations through aggregating the associated entities in the associated paths, instead of only the connected entities. The graph convolutional network consists of multiple graph convolution layers, and the message-passing process treats different entities discriminatorily according to the association strengths. For determining the plausibility of the knowledge, a scoring function is provided with the convolution operator. The experimental datasets are from a real grid company and contain four kinds of equipment. The experiments apply the proposed method to the equipment defects analysis, including the defect gradation, the coarse defect classification, and the fine defect classification. The proposed method is compared with some embedding methods. The experimental results verify that the proposed method outperforms the other methods for the real distribution network datasets, and the proposed method can analyze the defects effectively.

Automatic Outdoor Patrol Robot Based on Sensor Fusion and Face Recognition Methods

This study integrates path planning, fuzzy theory, neural networks, image processing, range sensors, webcam, global navigation satellite system (GNSS), and real-time kinematic (RTK) positioning system into an intelligent wheeled mobile robot (WMR) for outdoor patrolling. The robot system uses ultrasound sensors, laser sensors, and fuzzy controllers to detect obstacles and avoid them. The starting position and the goal position of the WMR in an outdoor environment are given by the GNSS RTK positioning system. Real-time position correction of the robot is performed through the differential global positioning system. The robot system applies the ant algorithm and the Dijkstra algorithm to find the shortest path for patrol tasks. The convolutional neural network image processing is utilized to identify intruders that are appearing in the patrol path. When the WMR detects an intruder by the face detection and recognition methods, the robot captures the photo of this person by the webcam and acquires the location information of this person by the RTK positioning system. Then the WMR sends the location and photo of the intruder to the control center by Wi-Fi and asks for help.

Design of Precise Delivery and Recovery System of Unmanned Aerial Vehicle Equipped with Line Patrol Robot

Based on the actual inspection requirements, this paper fully studies the application scenarios of UAV and robot inspection, and develops an intelligent UAV line inspection robot composite system based on UAV, robot, edge computing, machine vision, artificial intelligence and other technologies, which plays a significant role in promoting the application practice of UAV and robot inspection technology.

Rinegan: A Scalable Image Processing Architecture for Large Scale Surveillance Applications.

Image processing is widely used in intelligent robots, significantly improving the surveillance capabilities of smart buildings, industrial parks, and border ports. However, relying on the camera installed in a single robot is not enough since it only provides a narrow field of view as well as limited processing performance. Specially, a target person such as the suspect may appear anywhere and tracking the suspect in such a large-scale scene requires cooperation between fixed cameras and patrol robots. This induces a significant surge in demand for data, computing resources, as well as networking infrastructures. In this work, we develop a scalable architecture to optimize image processing efficacy and response rate for visual ability. In this architecture, the lightweight pre-process and object detection functions are deployed on the gateway-side to minimize the bandwidth consumption. Cloud-side servers receive solely the recognized data rather than entire image or video streams to identify specific suspect. Then the cloud-side sends the information to the robot, and the robot completes the corresponding tracking task. All these functions are implemented and orchestrated based on micro-service architecture to improve the flexibility. We implement a prototype system, called Rinegan, and evaluate it in an in-lab testing environment. The result shows that Rinegan is able to improve the effectiveness and efficacy of image processing.

Design of Intelligent Early Warning Robot System for Warehouse Autonomous Patrol based on Scale Estimation Algorithm of Probability Theory

In view of the backward monitoring status in the traditional safety management of hazardous chemicals warehouse, a patrol robot for hazardous chemicals warehouse with autonomous patrol stereo detection, intelligent early warning and remote monitoring functions is designed, and the design process of related functions of autonomous patrol navigation subsystem, environmental detection subsystem and remote monitoring subsystem is discussed from two aspects of hardware and software. Because both the camera and the target may be in motion, it is necessary to consider the scale change of target influence caused by relative distance. In order to overcome the shortcoming of invariable scale in Mean Shift algorithm, this paper adopts a scale adaptive method to estimate the size of the target, adjusts the parameters of kernel function by using ellipse feature, and takes the parameters of the current frame as the initial data of the next frame, so as to achieve the effect of adaptively adjusting the target scale. The test results show that the system can effectively traverse the map of hazardous chemicals warehouse, accurately assess the leakage of hazardous chemicals and the environmental safety of warehouse, and give early warning. The remote monitoring platform greatly facilitates the remote management of the warehouse.

Review of New Trends and Challenges of Android-Based Home Security Robot

Home security by itself is very costly and not many people, especially those who are living in rural or sub urban areas, are able to afford it. Hired security guards or even guard dogs are also slowly losing their usability and credibility in an age where burglars are sneakier, stealthier, and more technologically equipped to disable alarm systems. The Security Patrolling Robot vehicle moves in a set path while keeping its “eyes” peeled for any non-recognized intruders. The planned framework will utilize an infrared-based sensor following framework for supervision of various designated zones. The development of a robot is additionally controlled consequently through obstacle recognizing sensors to stay away from to avoid unnecessary collisions and possible damage to the robot. The working principle of the robot is that it screens every zone to recognize any interruptions, which is achieved utilizing a camera which is mounted on top of the robot to record the surveillance for the viewing of the client. The principal goal of this undertaking is to recognize the worrying exercises in the regions where the human presence cannot be detected. This robot will become a great contribution to the community. It is easily operated and can be handled by senior citizens, who may or may not be proficient with technology.

Women safety night patrolling robot using iot

India's greatest threat is the safety of its women. Women do not feel safe in a variety of situations. This needs to be addressed as quickly as feasible. Technology evolves and develops on a daily basis, affecting how people live. As a result, the focus of this article is on modernising the technological framework in order to strengthen women's safety mechanisms. We introduce a new security method in this study to protect women during unusual behaviours. A new security technique based on a patrolling robot using the Raspberry Pi has been proposed. In this case, a night vision camera can be employed to secure any location. Various machine learning methods are applied to improve the classifier's accuracy. The findings suggest that the suggested method outperforms existing methods.

Research on multi feature fusion perception technology of mine fire based on inspection robot

In order to improve the real-time and accuracy of mine fire identification, based on the summary of the existing fire monitoring methods, a mine fire early warning model based on patrol robot and multi-sensor and image recognition is proposed. The video image is obtained by using the high-definition camera of inspection robot, and the flame suspected area of video image is obtained by combining frame difference and background difference after filtering. The improved three scale analytic hierarchy process is used to fuse the circularity feature, sharp angle feature, centroid movement feature, color feature and area growth feature of the flame suspected area image to obtain the flame result. The three-frame difference method is used to binarize the smoke video image, and the smoke texture feature, smoke motion feature and smoke color feature are extracted to recognize the smoke image. The improved three-scale analytic hierarchy process (AHP) is used to fuse and judge the smoke in the combination of infrared thermal imager temperature measurement and smoke sensor. Inspection robot uses unified detection standards and intelligent analysis technology to sense mine smoke and flame. This method can improve the real-time, accuracy and robustness of mine fire perception.

Mobile Sentry Robot for Laboratory Safety Inspection Based on Machine Vision and Infrared Thermal Imaging Detection

As is known to all, university chemistry laboratories have always been considered as dangerous areas on campus, especially pharmaceutical chemistry laboratories, which are often the focus of pharmaceutical laboratory management. The innovation in safety management mode can effectively reduce or even eliminate security risks and reduce the risk of safety accidents. With the rapid development of artificial intelligence, robots have been introduced into all walks of life, replacing humans to complete part of the safety inspection. One of the core tasks of laboratory safety inspection is to find the source of fire and take measures to extinguish the fire in time. The intention of this paper is to design a sentry robot, which could take appropriate emergency measures to detect dangerous situations, including sounding alarms and carrying out firefighting, in order to minimize accidents and economic losses, based on machine vision for periodic inspection of laboratories with a temperature thermal imager to scan surroundings at the same time. The preliminary research results show that based on the machine vision, the robot identified the fire source and adjusted the spray device to carry out the fire-extinguishing experiment successfully. The development and utilization of laboratory patrol robots can share the work of laboratory patrol to a certain extent, guarantee the safety of experimental teaching in colleges and universities, and have room for sustainable research.

Visual inspection for transformer insulation defects by a patrol robot fish based on deep learning

In order to effectively detect the internal insulation defects of large transformers, a miniature patrol robot fish is designed and used to observe the pressboard inside the transformer, which can visually inspect the insulation condition of the pressboard. In the process of visual inspection by a patrol robot fish, insulation defects observed in the transformer (such as discharge carbon marks, pressboard cracks, etc.) are small size, indistinctive colour contrast and different shapes. The key purpose of the patrol robot fish is to identify the defect targets and defect types intelligently and quickly from the images taken by the camera on the fish. Considering that there are abundant insulation defects in transformers and a lack of samples for learning, a vision detection method based on deep learning network is proposed in this study. The proposed method integrates the variable autoencoder into the traditional Faster-RCNN target detection network and constructs an improved Faster-RCNN that enhances feature extraction. This method expands the small-scale training sample set and improves the generalisation performance of the model. In order to verify the effectiveness of the proposed method, the improved network is trained and tested, and the test results show that the improved network training model can accurately identify and mark the carbon marks on the pressboard surface.

Research on fire inspection robot based on computer vision

With the rapid development of robots in the world, robots in various application scenarios have come into people's sight, especially various inspection robots appear in people's lives. At present, the application of fire patrol robots is more and more. Now, the fire patrol robots used in the warehouse are all land patrol robots. When they work on the designated route, they have high requirements for the environment. When there is a complex environment with uneven ground, it will have a great impact on the robot. This rail mounted fire inspection robot is running on the track, and the platform can be inspected without dead angle. It has low environmental requirements, high inspection efficiency, real-time alarm, and uses computer vision platform to track, which improves the accuracy of fire source identification. In this design, STM32F4 single chip microcomputer is used as the main control of the whole system, which is completed with a number of servo motors, alarms and other hardware, and computer vision and robot main control data interaction.

Impressions of outdoor patrol robots for creating safe and secure places

Impressions of outdoor patrol robots for creating safe and secure places

3-D Ultrasonic Localization of Transformer Patrol Robot Based on EMD and PHAT-β Algorithms

This article is concerned with a hyperboloid 3-D localization method of the transformer patrol robot based on the empirical mode decomposition (EMD) of ultrasonic signals and phase transform-beta (PHAT- $\beta $ ) generalized cross correlation (GCC). A linear sensor array is first employed to receive ultrasonic signals emitted by robot. The received signals are decomposed into several intrinsic mode functions (IMFs) of different scales by EMD. According to the correlation degree between IMFs and the received signals, the IMFs with large correlation coefficients are subsequently selected to reconstruct the signals. Furthermore, the PHAT- $\beta $ GCC algorithm is used to estimate the time delay between the different sensors. Finally, multiple hyperboloids formed by sensors are intersected to accurately localize the robot. The simulation and experimental results demonstrate that the maximum relative errors in three typical positions are 5.6%, 4.8%, and 5.2%. The maximum localization error is less than 15 mm in a laboratory scale model with a size of length 800 mm, width 450 mm, and height 500 mm.

Development of functions for teleoperated patrolling robot

Development of functions for teleoperated patrolling robot