Benchmarking YOLOv5 models for improved human detection in search and rescue missions

Drone or unmanned aerial vehicle(UAV)technology has undergone significant changes.The technology allows UAV to carry out a wide range of tasks with an increasing level of sophistication,since drones can cover a large area with cameras.Meanwhile,the increasing number of computer vision applications utilizing deep learning provides a unique insight into such applications.The primary target in UAV-based detection applications is humans,yet aerial recordings are not included in the massive datasets used to train object detectors,which makes it necessary to gather the model data from such platforms.You only look once(YOLO)version 4,RetinaNet,faster region-based convolutional neural network(R-CNN),and cascade R-CNN are several well-known detectors that have been studied in the past using a variety of datasets to replicate rescue scenes.Here,we used the search and rescue(SAR)dataset to train the you only look once version 5(YOLOv5)algorithm to validate its speed,accuracy,and low false detection rate.In comparison to YOLOv4 and R-CNN,the highest mean average accuracy of 96.9%is obtained by YOLOv5.For comparison,experimental findings utilizing the SAR and the human rescue imaging database on land(HERIDAL)datasets are presented.The results show that the YOLOv5-based approach is the most successful human detection model for SAR missions.

Assessment of Meteorological Threats to the Coordinated Search and Rescue of Unmanned/Manned Aircraft

The architecture and working principle of coordinated search and rescue system of unmanned/manned aircraft,which is composed of manned/unmanned aircraft and manned aircraft,were first introduced,and they can cooperate with each other to complete a search and rescue task.Secondly,a threat assessment method based on meteorological data was proposed,and potential meteorological threats,such as storms and rainfall,can be predicted by collecting and analyzing meteorological data.Finally,an experiment was carried out to evaluate the performance of the proposed method in different scenarios.The experimental results show that the coordinated search and rescue system of unmanned/manned aircraft can be used to effectively assess meteorological threats and provide accurate search and rescue guidance.

Structural Design Study of Air-Dropped Unmanned Maritime Mobile Search and Rescue Platforms

In order to improve the efficiency and safety of search and rescue(SAR)at sea,this paper proposes a kind of emergency rapid rescue unmanned craft(air-dropped unmanned maritime motorized search and rescue platform)that can be delivered by a large transport aircraft.This paper studies the structural design scheme of the platform,and the main scale of the platform,the choice of power system and the impact resistance performance are considered in the design process to ensure its rapid response and effective rescue capability under complex sea conditions.Simulation results show that the platform can withstand the impact of air injection into the water and the shipboard equipment can operate normally under the impact load,thus verifying the feasibility and safety of the design.This study serves to improve the maritime search and rescue system and enhance the oceanic emergency response capability.

Search and Rescue Optimization with Machine Learning Enabled Cybersecurity Model

Presently,smart cities play a vital role to enhance the quality of living among human beings in several ways such as online shopping,e-learning,ehealthcare,etc.Despite the benefits of advanced technologies,issues are also existed from the transformation of the physical word into digital word,particularly in online social networks(OSN).Cyberbullying(CB)is a major problem in OSN which needs to be addressed by the use of automated natural language processing(NLP)and machine learning(ML)approaches.This article devises a novel search and rescue optimization with machine learning enabled cybersecurity model for online social networks,named SRO-MLCOSN model.The presented SRO-MLCOSN model focuses on the identification of CB that occurred in social networking sites.The SRO-MLCOSN model initially employs Glove technique for word embedding process.Besides,a multiclass-weighted kernel extreme learning machine(M-WKELM)model is utilized for effectual identification and categorization of CB.Finally,Search and Rescue Optimization(SRO)algorithm is exploited to fine tune the parameters involved in the M-WKELM model.The experimental validation of the SRO-MLCOSN model on the benchmark dataset reported significant outcomes over the other approaches with precision,recall,and F1-score of 96.24%,98.71%,and 97.46%respectively.

DEVELOPMENT OF A SHAPE-SHIFTING MOBILE ROBOT FOR URBAN SEARCH AND RESCUE

A portable shape-shifting mobile robot system named as Amoeba Ⅱ（A-Ⅱ） is developed for the urban search and rescue application. It is designed with three degrees of freedom and two tracked drive systems. This robot consists of two modular mobile units and a joint unit. The mobile unit is a tracked mechanism to enforce the propulsion of robot. And the joint unit can transform the robot shape to get high environment adaptation. A-Ⅱ robot can not only adapt to the environment but also change its body shape according to the locus space. It behaves two work states including the linear state （named as I state） and the parallel state （named as Ⅱ state）. With the linear state the robot can climb upstairs and go through narrow space such as the pipe, cave, etc. The parallel state enables the robot with high mobility on rough ground. Also, the joint unit can propel the robot to roll in sidewise direction. Two modular A-Ⅱ robots can be connected through jointing common interfaces on the joint unit to compose a stronger shape-shifting robot, which can transform the body into four wheels-driven vehicle. The experimental results validate the adaptation and mobility of A-Ⅱ robot.

Robot SLAM with Ad hoc wireless network adapted to search and rescue environments

An innovative multi-robot simultaneous localization and mapping(SLAM)is proposed based on a mobile Ad hoc local wireless sensor network(Ad-WSN).Multiple followed-robots equipped with the wireless link RS232/485module act as mobile nodes,with various on-board sensors,Tp-link wireless local area network cards,and Tp-link wireless routers.The master robot with embedded industrial PC and a complete robot control system autonomously performs the SLAM task by exchanging information with multiple followed-robots by using this self-organizing mobile wireless network.The PC on the remote console can monitor multi-robot SLAM on-site and provide direct motion control of the robots.This mobile Ad-WSN complements an environment devoid of usual GPS signals for the robots performing SLAM task in search and rescue environments.In post-disaster areas,the network is usually absent or variable and the site scene is cluttered with obstacles.To adapt to such harsh situations,the proposed self-organizing mobile Ad-WSN enables robots to complete the SLAM process while improving the performances of object of interest identification and exploration area coverage.The information of localization and mapping can communicate freely among multiple robots and remote PC control center via this mobile Ad-WSN.Therefore,the autonomous master robot runs SLAM algorithms while exchanging information with multiple followed-robots and with the remote PC control center via this local WSN environment.Simulations and experiments validate the improved performances of the exploration area coverage,object marked,and loop closure,which are adapted to search and rescue post-disaster cluttered environments.

AModified Search and Rescue Optimization Based Node Localization Technique inWSN

Wireless sensor network(WSN)is an emerging technology which find useful in several application areas such as healthcare,environmentalmonitoring,border surveillance,etc.Several issues that exist in the designing of WSN are node localization,coverage,energy efficiency,security,and so on.In spite of the issues,node localization is considered an important issue,which intends to calculate the coordinate points of unknown nodes with the assistance of anchors.The efficiency of the WSN can be considerably influenced by the node localization accuracy.Therefore,this paper presents a modified search and rescue optimization based node localization technique(MSRONLT)forWSN.The major aim of theMSRO-NLT technique is to determine the positioning of the unknown nodes in theWSN.Since the traditional search and rescue optimization(SRO)algorithm suffers from the local optima problemwith an increase in number of iterations,MSRO algorithm is developed by the incorporation of chaotic maps to improvise the diversity of the technique.The application of the concept of chaotic map to the characteristics of the traditional SRO algorithm helps to achieve better exploration ability of the MSRO algorithm.In order to validate the effective node localization performance of the MSRO-NLT algorithm,a set of simulations were performed to highlight the supremacy of the presented model.A detailed comparative results analysis showcased the betterment of the MSRO-NLT technique over the other compared methods in terms of different measures.

Research Review on Marine Search and Rescue

Locating the marine target in a quick and precise way is the crucial point of implementing SAR （search and rescue） at sea, which involves aspects of developing SAR strategy and detects the marine targets. As the effect of marine target detection restricts the SAR result directly, the study has focused on reviewing the previous research about marine target detection, especially dim marine target detection. What＇s more, small target detection under complex sea status is one of the severe challenges which is research＇s hotspot and needs more endeavor. Current research results and future research directions are discussed in the paper. The findings can provide systematic view of implementing maritime search and rescue for field researchers and governors.

A scenario construction and similarity measurement method for navy combat search and rescue

Navy combat search and rescue(NCSAR) is an important component of the modern maritime warfare and the scenario of NCSAR is the basis for decision makers to rely on. According to the core elements in the NCSAR process, the NCSAR scenario structure is constructed from seven perspectives based on the multi-view architecture framework. According to the NCSAR scenarios evolution over time, the NCSAR scenario sequence is analyzed and modeled based on the concept lattice method. Then,the incremental construction algorithm of the NCSAR scenario sequence lattice is given. On this basis, the similarity measurement index of NCSAR scenarios is defined, and the similarity measurement model of NCSAR scenarios is proposed. Finally, the rationality of the method is verified by an example analysis. The NCSAR scenario and similarity measurement method proposed can provide scientific guidance for rapid making, dynamic adjustment and implementation of the NCSAR program, and thus improve the efficiency and effectiveness of NCSAR.

A Wireless Communication Method with Dynamic Adding Nodes for the Underground Search and Rescue Robot

A wireless communication method with dynamic adding nodes for Underground Search and Rescue robot is proposed： fix the address of the controller, add repeater nodes into the net dynamically, and shift the address of the mobile terminal. With this method, the Search and Rescue robot can reach the deeper place of a mine to help rescue and keep in touch with the controller through wireless communication in a single channel, even in a complex laneway where radio wave cannot go through the thick wall. The collision in the process of the two-way multi-hop communication in the single channel will also be resolved by the communication direction priority and response signal mechanism, to enhance the reliability of communication. Finally, a sample is designed and an experiment is conducted to verify the efficiency of the method.

Vision and Geolocation Data Combination for Precise Human Detection and Tracking in Search and Rescue Operations

In this paper, a study and evaluation of the combination of GPS/GNSS techniques and advanced image processing algorithms for distressed human detection, positioning and tracking, from a fully autonomous Unmanned Aerial Vehicle (UAV)-based rescue support system, are presented. In particular, the issue of human detection both on terrestrial and marine environment under several illumination and background conditions, as the human silhouette in water differs significantly from a terrestrial one, is addressed. A robust approach, including an adaptive distressed human detection algorithm running every N input image frames combined with a much faster tracking algorithm, is proposed. Real time or near-real-time distressed human detection rates achieved, using a single, low cost day/night NIR camera mounted onboard a fully autonomous UAV for Search and Rescue (SAR) operations. Moreover, the generation of our own dataset, for the image processing algorithms training is also presented. Details about both hardware and software configuration as well as the assessment of the proposed approach performance are fully discussed. Last, a comparison of the proposed approach to other human detection methods used in the literature is presented.

Detection of leaf folder and yellow stemborer moths in the paddy field using deep neural network with search and rescue optimization

In agriculture,insect pests must be identified at the initial stage of infestation to avoid their spread in the field.Leaf folders(cnaphalocrocis medinalis)and yellow stemborers(scirpophaga incertulas)are destructive pests of paddy crops,which are causing severe yield loss.Manual identification of insect pests in the crop is time-consuming,tedious,and ineffective.This paper focuses on a light trap based four-layer deep neural network with search and rescue optimization(DNN-SAR)method to identify leaf folders and yellow stemborers.Light traps are designed to lure the insects in the paddy field and the images of trapped insects are analyzed using the proposed detection method.In the DNN-SAR,images are contrastenhanced using deer hunting algorithm,impulse noise is removed with fast average group filter,and segmented using social ski-driver optimization.The search and rescue optimization algorithm is used for the selection of optimal weights in the deep neural network,which has improved the convergence rate,lowered the complexity of learning,and improved the accuracy of detection.The proposed method outperformed the existing methods and achieved 98.29%pest detection accuracy.

Evaluation method for helicopter maritime search and rescue response plan with uncertainty

Helicopter plays an increasingly significant role in Maritime Search and Rescue(MSAR),and it will perform MSAR mission based on response plans when an accident occurs.Thus the rationality of response plan determines the success of MSAR mission to a large extent.However,with the impact of many uncertainty factors,it is difficult to evaluate response plans comprehensively before performing them.Aiming at these problems,an evaluation framework of helicopter MSAR response plan named UMAD is proposed in this paper,which reveals the influence mechanism of uncertainty factors based on Multi-Agent method and analyzes the mission flow based on Discrete Event System(DEVS)method.Furthermore,the evaluation criterion and indicators of response plan are extracted from the aspects of safety and effectiveness.Meanwhile,the Monte Carlo method is adapted to calculate the probability distribution and robustness of response plan comprehensive result.Finally,in order to illustrate the validity of this method,it is discussed and verified by an application example of evaluating multiple response plans to the same MSAR scenario.The results show that this method can analyze the influence of uncertainty more systematically and optimize response plans more comprehensively.

Maritime Search and Rescue Networking Based on Multi-Agent Cooperative Communication

Rapid and effective maritime search and rescue operations become the important guarantee for the safety of maritime navigation.The existing maritime search and rescue networking and model have slow response speed and low efficiency.The distribution,synergy,parallelism,robustness and intelligence of unmanned surface vehicle(USV)and unmanned aerial vehicle(UAV)provide a new idea for the novel maritime search and rescue networking,in which multi-agent could be used to build a layered control network.In this paper,a novel rapid search and rescue system is proposed by utilizing the improved ant colony optimization and the independent calculation decision of the agents.The system adopts the edge computing,relies on the information sharing and the cooperative decision between the search and rescue agent groups.It achieves the independent synchronous search and rescue.At the same time,we use particle swarm optimization to intelligently schedule data packets during the rescue process to optimize network forwarding performance.Based on the distributed cluster control of USV and UAV,this paper combines edge computing,cooperative communication and centralized task allocation together to make decision for rescue.The simulation results show that our proposed schemes realize a significant improvement for maritime search and rescue.

Should Search and Rescue Operations Be Free of Charge?

On April 3, 39 teachers and students from the Beijing Institute of Technology (BIT)were trapped on the MaoerMountain in Fangshan District,a suburban area in Beijing. Morethan 300 persons,

An anti-collision algorithm for robotic search-and-rescue tasks in unknown dynamic environments

This paper deals with the search-and-rescue tasks of a mobile robot with multiple interesting targets in an unknown dynamic environment.The problem is challenging because the mobile robot needs to search for multiple targets while avoiding obstacles simultaneously.To ensure that the mobile robot avoids obstacles properly,we propose a mixed-strategy Nash equilibrium based Dyna-Q(MNDQ)algorithm.First,a multi-objective layered structure is introduced to simplify the representation of multiple objectives and reduce computational complexity.This structure divides the overall task into subtasks,including searching for targets and avoiding obstacles.Second,a risk-monitoring mechanism is proposed based on the relative positions of dynamic risks.This mechanism helps the robot avoid potential collisions and unnecessary detours.Then,to improve sampling efficiency,MNDQ is presented,which combines Dyna-Q and mixed-strategy Nash equilibrium.By using mixed-strategy Nash equilibrium,the agent makes decisions in the form of probabilities,maximizing the expected rewards and improving the overall performance of the Dyna-Q algorithm.Furthermore,a series of simulations are conducted to verify the effectiveness of the proposed method.The results show that MNDQ performs well and exhibits robustness,providing a competitive solution for future autonomous robot navigation tasks.

Machine-to-Machine Collaboration Utilizing Internet of Things and Machine Learning

Machine-to-Machine (M2M) collaboration opens new opportunities where systems can collaborate without any human intervention and solve engineering problems efficiently and effectively. M2M is widely used for various application areas. Through this reported project authors developed a M2M system where a drone and two ground vehicles collaborate through a base station to implement a system that can be utilized for an indoor search and rescue operation. The model training for drone flight paths achieves almost 100% accuracy. It was also observed that the accuracy of the model increased with more training samples. Both the drone flight path and ground vehicle navigation are controlled from the base station. Machine learning is utilized for modelling of drone’s flight path as well as for ground vehicle navigation through obstacles. The developed system was implemented on a field trial within a corridor of a building, and it was demonstrated successfully.

Efficient Deep Learning Framework for Fire Detection in Complex Surveillance Environment

To prevent economic,social,and ecological damage,fire detection and management at an early stage are significant yet challenging.Although computationally complex networks have been developed,attention has been largely focused on improving accuracy,rather than focusing on real-time fire detection.Hence,in this study,the authors present an efficient fire detection framework termed E-FireNet for real-time detection in a complex surveillance environment.The proposed model architecture is inspired by the VGG16 network,with significant modifications including the entire removal of Block-5 and tweaking of the convolutional layers of Block-4.This results in higher performance with a reduced number of parameters and inference time.Moreover,smaller convolutional kernels are utilized,which are particularly designed to obtain the optimal details from input images,with numerous channels to assist in feature discrimination.In E-FireNet,three steps are involved:preprocessing of collected data,detection of fires using the proposed technique,and,if there is a fire,alarms are generated and transmitted to law enforcement,healthcare,and management departments.Moreover,E-FireNet achieves 0.98 accuracy,1 precision,0.99 recall,and 0.99 F1-score.A comprehensive investigation of various Convolutional Neural Network(CNN)models is conducted using the newly created Fire Surveillance SV-Fire dataset.The empirical results and comparison of numerous parameters establish that the proposed model shows convincing performance in terms of accuracy,model size,and execution time.

Helicopter maritime search area planning based on a minimum bounding rectangle and K-means clustering

Helicopters are widely used in maritime Search and Rescue(SAR) missions. To ensure the success of SAR missions, search areas need to be carefully planned. With the development of computer technology and weather forecast technology, the survivors’ drift trajectories can be predicted more precisely, which strongly supports the planning of search areas for the rescue helicopter. However, the methods used to determine the search area based on the predicted drift trajectories are mainly derived from the continuous expansion of the area with the highest Probability of Containment(POC), which may lead to local optimal solutions and a decrease in the Probability of Success(POS), especially when there are several subareas with a high POC. To address this problem, this paper proposes a method based on a Minimum Bounding Rectangle and Kmeans clustering(MBRK). A silhouette coefficient is adopted to analyze the distribution of the survivors’ probable locations, which are divided into multiple clusters with K-means clustering. Then,probability maps are generated based on the minimum bounding rectangle of each cluster. By adding or subtracting one row or column of cells or shifting the planned search area, 12 search methods are used to generate the optimal search area starting from the cell with the highest POC in each probability map. Taking a real case as an example, the simulation experiment results show that the POS values obtained by the MBRK method are higher than those obtained by other methods,which proves that the MBRK method can effectively support the planning of search areas and that K-means clustering improves the POS of search plans.

Wilderness medicine

BACKGROUND: Human activity in wilderness areas has increased globally in recent decades, leading to increased risk of injury and illness. Wilderness medicine has developed in response to both need and interest.METHODS: The field of wilderness medicine encompasses many areas of interest. Some focus on special circumstances(such as avalanches) while others have a broader scope(such as trauma care). Several core areas of key interest within wilderness medicine are discussed in this study.RESULTS: Wilderness medicine is characterized by remote and improvised care of patients with routine or exotic illnesses or trauma, limited resources and manpower, and delayed evacuation to definitive care. Wilderness medicine is developing rapidly and draws from the breadth of medical and surgical subspecialties as well as the technical fields of mountaineering, climbing, and diving. Research, epidemiology, and evidence-based guidelines are evolving. A hallmark of this field is injury prevention and risk mitigation. The range of topics encompasses high-altitude cerebral edema, decompression sickness, snake envenomation, lightning injury, extremity trauma, and gastroenteritis. Several professional societies, academic fellowships, and training organizations offer education and resources for laypeople and health care professionals.CONCLUSIONS: The future of wilderness medicine is unfolding on multiple fronts: education, research, training, technology, communications, and environment. Although wilderness medicine research is technically difficult to perform, it is essential to deepening our understanding of the contribution of specific techniques in achieving improvements in clinical outcomes.