Showing papers on "Obstacle published in 2022"
01 Mar 2022
TL;DR: The authors suggest language of evidence that allows for a more nuanced approach to communicate scientific findings as a simple and intuitive alternative to statistical significance testing, and provide examples for rewriting results sections in research papers accordingly.
Abstract: Despite much criticism, black-or-white null-hypothesis significance testing with an arbitrary P-value cutoff still is the standard way to report scientific findings. One obstacle to progress is likely a lack of knowledge about suitable alternatives. Here, we suggest language of evidence that allows for a more nuanced approach to communicate scientific findings as a simple and intuitive alternative to statistical significance testing. We provide examples for rewriting results sections in research papers accordingly. Language of evidence has previously been suggested in medical statistics, and it is consistent with reporting approaches of international research networks, like the Intergovernmental Panel on Climate Change, for example. Instead of re-inventing the wheel, ecology and evolution might benefit from adopting some of the 'good practices' that exist in other fields.
149 citations
TL;DR: A large-scale dataset of object detection in aerial images (DOTA) is presented in this article , which contains 1,793,658 object instances of 18 categories of oriented-bounding-box annotations collected from 11,268 aerial images.
Abstract: In he past decade, object detection has achieved significant progress in natural images but not in aerial images, due to the massive variations in the scale and orientation of objects caused by the bird's-eye view of aerial images. More importantly, the lack of large-scale benchmarks has become a major obstacle to the development of object detection in aerial images (ODAI). In this paper, we present a large-scale Dataset of Object deTection in Aerial images (DOTA) and comprehensive baselines for ODAI. The proposed DOTA dataset contains 1,793,658 object instances of 18 categories of oriented-bounding-box annotations collected from 11,268 aerial images. Based on this large-scale and well-annotated dataset, we build baselines covering 10 state-of-the-art algorithms with over 70 configurations, where the speed and accuracy performances of each model have been evaluated. Furthermore, we provide a code library for ODAI and build a website for evaluating different algorithms. Previous challenges run on DOTA have attracted more than 1300 teams worldwide. We believe that the expanded large-scale DOTA dataset, the extensive baselines, the code library and the challenges can facilitate the designs of robust algorithms and reproducible research on the problem of object detection in aerial images.
61 citations
TL;DR: Novel flexible shape-adaptive selection (SA-S) and shape- Adaptive measurement (SA -M) strategies for oriented object detection, which comprise an SA-S strategy for sample selection and SA-M strategy for the quality estimation of positive samples are proposed.
Abstract: The development of detection methods for oriented object detection remains a challenging task. A considerable obstacle is the wide variation in the shape (e.g., aspect ratio) of objects. Sample selection in general object detection has been widely studied as it plays a crucial role in the performance of the detection method and has achieved great progress. However, existing sample selection strategies still overlook some issues: (1) most of them ignore the object shape information; (2) they do not make a potential distinction between selected positive samples; and (3) some of them can only be applied to either anchor-free or anchor-based methods and cannot be used for both of them simultaneously. In this paper, we propose novel flexible shape-adaptive selection (SA-S) and shape-adaptive measurement (SA-M) strategies for oriented object detection, which comprise an SA-S strategy for sample selection and SA-M strategy for the quality estimation of positive samples. Specifically, the SA-S strategy dynamically selects samples according to the shape information and characteristics distribution of objects. The SA-M strategy measures the localization potential and adds quality information on the selected positive samples. The experimental results on both anchor-free and anchor-based baselines and four publicly available oriented datasets (DOTA, HRSC2016, UCAS-AOD, and ICDAR2015) demonstrate the effectiveness of the proposed method.
43 citations
TL;DR: Zhang et al. as discussed by the authors proposed a path planning with dynamic obstacle avoidance method of the manipulator based on a deep reinforcement learning algorithm soft actor-critic (SAC) to avoid the moving obstacle in the environment and make real-time planning.
38 citations
TL;DR: The goal is to structure a modest, secure, wearable, and versatile framework for visually impaired to help them in their daily routines, and detects the obstructions with a notably high efficiency.
Abstract: Recent progress in innovation is making the life prosper, simpler and easier for common individual. The World Health Organization (WHO) statistics indicate that a large amount of people experience visual losses, because of which they encounter many difficulties in everyday jobs. Hence, our goal is to structure a modest, secure, wearable, and versatile framework for visually impaired to help them in their daily routines. The proposed methodology utilizes Raspberry-Pi 4B, Camera, Ultrasonic Sensor and Arduino, mounted on the stick of the individual. We take pictures of the scene and afterwards pre-process these pictures with the help of Viola Jones and TensorFlow Object Detection algorithm. The said techniques are used to detect objects. We also used an ultrasonic sensor mounted on a servomotor to measure the distance between the blind person and obstacles. The presented research utilizes simple calculations for its execution, and detects the obstructions with a notably high efficiency. When contrasted with different frameworks, this framework is a minimal effort, convenient, and simple to wear.
35 citations
TL;DR: The real-time neural network detector architecture You Only Look Once, the latest version (YOLOv4), is investigated and it is demonstrated that this detector can be adapted to multispectral pedestrian detection and can achieve accuracy on par with the state-of-the-art while being highly computationally efficient, thereby supporting low-latency decision making.
Abstract: Detecting pedestrians in autonomous driving is a safety-critical task, and the decision to avoid a a person has to be made with minimal latency. Multispectral approaches that combine RGB and thermal images are researched extensively, as they make it possible to gain robustness under varying illumination and weather conditions. State-of-the-art solutions employing deep neural networks offer high accuracy of pedestrian detection. However, the literature is short of works that evaluate multispectral pedestrian detection with respect to its feasibility in obstacle avoidance scenarios, taking into account the motion of the vehicle. Therefore, we investigated the real-time neural network detector architecture You Only Look Once, the latest version (YOLOv4), and demonstrate that this detector can be adapted to multispectral pedestrian detection. It can achieve accuracy on par with the state-of-the-art while being highly computationally efficient, thereby supporting low-latency decision making. The results achieved on the KAIST dataset were evaluated from the perspective of automotive applications, where low latency and a low number of false negatives are critical parameters. The middle fusion approach to YOLOv4 in its Tiny variant achieved the best accuracy to computational efficiency trade-off among the evaluated architectures.
30 citations
TL;DR: This letter proposes a distributed approach for multi-robot navigation which combines the concept of reciprocal velocity obstacle (RVO) and the scheme of deep reinforcement learning (DRL) to solve the reciprocal collision avoidance problem under limited information.
Abstract: The challenges to solving the collision avoidance problem lie in adaptively choosing optimal robot velocities in complex scenarios full of interactive obstacles. In this letter, we propose a distributed approach for multi-robot navigation which combines the concept of reciprocal velocity obstacle (RVO) and the scheme of deep reinforcement learning (DRL) to solve the reciprocal collision avoidance problem under limited information. The novelty of this work is threefold: (1) using a set of sequential VO and RVO vectors to represent the interactive environmental states of static and dynamic obstacles, respectively; (2) developing a bidirectional recurrent module based neural network, which maps the states of a varying number of surrounding obstacles to the actions directly; (3) developing a RVO area and expected collision time based reward function to encourage reciprocal collision avoidance behaviors and trade off between collision risk and travel time. The proposed policy is trained through simulated scenarios and updated by the actor-critic based DRL algorithm. We validate the policy in complex environments with various numbers of differential drive robots and obstacles. The experiment results demonstrate that our approach outperforms the state-of-art methods and other learning based approaches in terms of the success rate, travel time, and average speed.
30 citations
TL;DR: In this paper , an uncertain moving obstacle avoidance method based on the improved velocity obstacle method was designed to reduce the distance and time of obstacle avoidance, and a series of experiments were carried out in the pool that validates the proposed methods are also presented.
30 citations
TL;DR: In this article , the authors investigated the effect of obstacle height on the performance of a proton exchange membrane fuel cell (PEMFC) with tri-serpentine channel and found that the presence of obstacle could enhance the transfer of reactant gas towards GDL and widen the working range of current density.
28 citations
TL;DR: This research introduces “Vision Navigator,” a novel framework for assisting visually impaired users in obstacle analysis and tracking so that they can move independently.
Abstract: Vision impairment is a major challenge faced by humanity on a large scale throughout the world. Affected people find independently navigating and detecting obstacles extremely tedious. Thus, a potential solution for accurately detecting obstacles requires an integrated deployment of the Internet of Things and predictive analytics. This research introduces “Vision Navigator,” a novel framework for assisting visually impaired users in obstacle analysis and tracking so that they can move independently. An intelligent stick named “Smart-fold Cane” and sensor-equipped shoes called “Smart-alert Walker” are the main constituents of our proposed model. For object detection and classification, the stick uses a single-shot detection (SSD) mechanism, which is followed by frame generation using the recurrent neural network (RNN) model. Smart-alert Walker is a lightweight shoe that acts as an emergency unit that notifies the user regarding the presence of any obstacle within a short distance range. This intelligent obstacle detection model using the SSD-RNN approach was deployed in real time and its performance was validated in indoor and outdoor environments. The SSD-RNN model computed an optimum accuracy of 95.06% and 87.68% indoors and outdoors, respectively. The model was also evaluated in the context of users’ distance from obstacles. The proposed SSD-RNN model had an accuracy rate of 96.4% and 86.8% for close and distant obstacles, respectively, outperforming other models. Execution time for the SSD-RNN model was 4.82 s with the highest mean accuracy rate of 95.54% considering all common obstacles.
28 citations
01 Jan 2022
TL;DR: In this paper, a target-tracking controller for the UAVs is developed to meet the requirements of payload dynamics and obstacle avoidance, and the authors also present simulation results to demonstrate the benefits of the proposed problem formulation for a multiobstacle environment.
Abstract: Control barrier functions have been widely studied and applied to safety-critical systems, including multi-agent obstacle avoidance problems. In this work, we apply control barrier functions to a collaborative transportation problem involving two unmanned aerial vehicles (UAVs) moving a payload around obstacles as they deliver it to a target location. We develop a target-tracking controller for the UAVs, which is constrained to meet the requirements of payload dynamics and obstacle avoidance. We also present simulation results to demonstrate the benefits of the proposed problem formulation for a multi-obstacle environment.
TL;DR: In this paper , the authors propose a solution to address the concerns of both developing countries and life scientists by ensuring international benefit sharing from sequence data without jeopardizing open sharing is a major obstacle for the Convention on Biological Diversity and other UN negotiations.
Abstract: Open access to sequence data is a cornerstone of biology and biodiversity research, but has created tension under the United Nations Convention on Biological Diversity (CBD). Policy decisions could compromise research and development, unless a practical multilateral solution is implemented. Ensuring international benefit-sharing from sequence data without jeopardising open sharing is a major obstacle for the Convention on Biological Diversity and other UN negotiations. Here, the authors propose a solution to address the concerns of both developing countries and life scientists.
31 Mar 2022
TL;DR: In this paper , a mixed boundary value problem with a nonhomogeneous, nonlinear differential operator (called a double phase operator), a nonlinear convection term (a reaction term depending on the gradient), three multivalued terms, and an implicit obstacle constraint is considered.
Abstract: In this paper we consider a mixed boundary value problem with a nonhomogeneous, nonlinear differential operator (called a double phase operator), a nonlinear convection term (a reaction term depending on the gradient), three multivalued terms, and an implicit obstacle constraint. Under very general assumptions on the data, we prove that the solution set of such an implicit obstacle problem is nonempty (so there is at least one solution) and weakly compact. The proof of our main result uses the Kakutani-Ky Fan fixed point theorem for multivalued operators along with the theory of nonsmooth analysis and variational methods for pseudomonotone operators.
TL;DR: A detailed survey on mmWave radar and vision fusion based obstacle detection methods is presented in this paper , where the authors classify the fusion methods into data level, decision level, and feature level fusion methods.
Abstract: With autonomous driving developing in a booming stage, accurate object detection in complex scenarios attract wide attention to ensure the safety of autonomous driving. Millimeter wave (mmWave) radar and vision fusion is a mainstream solution for accurate obstacle detection. This article presents a detailed survey on mmWave radar and vision fusion based obstacle detection methods. First, we introduce the tasks, evaluation criteria, and datasets of object detection for autonomous driving. The process of mmWave radar and vision fusion is then divided into three parts: sensor deployment, sensor calibration, and sensor fusion, which are reviewed comprehensively. Specifically, we classify the fusion methods into data level, decision level, and feature level fusion methods. In addition, we introduce three-dimensional(3D) object detection, the fusion of lidar and vision in autonomous driving and multimodal information fusion, which are promising for the future. Finally, we summarize this article.
TL;DR: In this paper , the authors present a framework for the safety-critical control of robotic systems, when safety is defined on safe regions in the configuration space, and synthesize a safe velocity based on control barrier function theory without relying on a high-fidelity dynamical model of the robot.
Abstract: This paper presents a framework for the safety-critical control of robotic systems, when safety is defined on safe regions in the configuration space. To maintain safety, we synthesize a safe velocity based on control barrier function theory without relying on a -- potentially complicated -- high-fidelity dynamical model of the robot. Then, we track the safe velocity with a tracking controller. This culminates in model-free safety critical control. We prove theoretical safety guarantees for the proposed method. Finally, we demonstrate that this approach is application-agnostic. We execute an obstacle avoidance task with a Segway in high-fidelity simulation, as well as with a Drone and a Quadruped in hardware experiments.
27 Jun 2022
TL;DR: This work has proposed a robust LoG-based deep neural network for object detection and tracking, which has great advantages in robustness to object scale and illuminations compared with typical deep network-based approaches and improved based on the original Kalman filter.
Abstract: The vision of unmanned aerial vehicles is very significant for UAV-related applications such as search and rescue, landing on a moving platform, etc. In this work, we have developed an integrated system for the UAV landing on the moving platform, and the UAV object detection with tracking in the complicated environment. Firstly, we have proposed a robust LoG-based deep neural network for object detection and tracking, which has great advantages in robustness to object scale and illuminations compared with typical deep network-based approaches. Then, we have also improved based on the original Kalman filter and designed an iterative multi-model-based filter to tackle the problem of unknown dynamics in real circumstances of motion estimations. Next, we have implemented the whole system and do ROS Gazebo-based testing in two complicated circumstances to verify the effectiveness of our design. Finally, we have deployed the proposed detection, tracking, and motion estimation strategies into real applications to do UAV tracking of a pillar and do obstacle avoidance. It is demonstrated that our system shows great accuracy and robustness in real applications.
TL;DR: In this paper , the effect of obstacle on the syngas (H2/CO) explosion characteristics under lean-fuel conditions was investigated, and the relationship between the flame propagation characteristics and explosion intensity, and effects of Syngas concentration, obstacle number and blocking rate on explosion parameters were obtained.
01 Jan 2022
TL;DR: In this article, an adaptive coverage control strategy for multi-agent systems with obstacle avoidance in the presence of actuator faults and time-varying uncertainties is presented based on a leader-follower approach.
Abstract: This letter presents an adaptive coverage control strategy for multi-agent systems with obstacle avoidance in the presence of actuator faults and time-varying uncertainties. The strategy is based on a leader-follower approach. Assuming that the motion of the leader is given, one distributes the followers within the leader’s obstacle-free sensing range so that collisions with obstacles can be avoided. An optimized distribution is achieved through the Centroidal Voronoi Tessellation (CVT) and a function approximation technique based immersion and invariance (FATII) coverage controller is constructed to realize the CVT. The stability of the FATII coverage controller is established and its validity is tested by simulations.
TL;DR: A comprehensive survey on carpooling in autonomous and connected vehicles is presented in this article , which covers architecture, components, and solutions, including scheduling, matching, mobility, pricing models of car-sharing.
Abstract: Owing to the advancements in communication and computation technologies, the dream of commercialized connected and autonomous cars is becoming a reality. However, among other challenges such as environmental pollution, cost, maintenance, security, and privacy, the ownership of vehicles (especially for Autonomous Vehicles) is the major obstacle in the realization of this technology at the commercial level. Furthermore, the business model of pay-as-you-go type services further attracts the consumer, because there is no need for upfront investment. In this vein, the idea of car-sharing (aka carpooling) is getting ground due to, at least in part, its simplicity, cost-effectiveness, and affordable choice of transportation. Carpooling systems are still in their infancy and face challenges such as scheduling, matching passengers interests, business model, security, privacy, and communication. To date, a plethora of research work has already been done covering different aspects of carpooling services (ranging from applications to communication and technologies); however, there is still a lack of a holistic, comprehensive survey that can be a one-stop-shop for the researchers in this area to (i) find all the relevant information and (ii) identify the future research directions. To fill these research challenges, this article provides a comprehensive survey on carpooling in autonomous and connected vehicles and covers architecture, components, and solutions, including scheduling, matching, mobility, pricing models of carpooling. We also discuss the current challenges in carpooling and identify future research directions. This survey is aimed to spur further discussion among the research community for the effective realization of carpooling.
TL;DR: Wang et al. as mentioned in this paper proposed a Deep Reinforcement Learning (DRL)-based method that enables unmanned aerial vehicles (UAVs) to execute navigation tasks in multi-obstacle environments with randomness and dynamics.
TL;DR: In this article , an experimental study of the flame propagation behavior influenced by quadrangular, cylindrical, and triangular obstacles was executed in a 530 mm × 82 mm × 12.5 mm pipe, and the results confirm the understanding that obstacles with a tip promote the generation of flow instability and produce a more intense burning behavior of a flame.
Abstract: The safe design and operation of clean fuels like hydrogen require a detailed understanding of their explosion characteristics. An experimental study of the flame propagation behavior influenced by quadrangular, cylindrical, and triangular obstacles was executed in a 530 mm × 82 mm × 82 mm pipe. The results confirm the understanding that obstacles with a tip promote the generation of flow instability and produce a more intense burning behavior of the flame. The shear layers shed fewer larger vortices after the quadrangular obstacle; however, these vortices can be dislodged to form smaller and more vortices after passing through the triangular obstacles. The shear layer has weak shedding properties behind obstacles with curved edges. In the flame propagation process, the quadrangular obstacles have a more obvious promoting effect of the initial explosion, but the degree is weaker than with the triangular obstacles. The effect of the quadrangular obstacles on flame velocity is mainly influenced by gas flow at the flame front. In triangular obstacles, the shear layer became prominent in the later stage of the explosion process and this contributes to enhancing the flame velocity and overpressure.
TL;DR: Me Mask R-CNN as mentioned in this paper was proposed to improve the accuracy of active identification of train obstacles by using SSwin-Le Transformer as the feature extraction network and ME-PAPN was used as feature fusion network, which achieved 91.3 % mAP with an average detection time of 4.2 FPS.
03 Jan 2022
TL;DR: In this article , an improved version of stochastic exploration namely Frequency Modified Hybrid Whale Optimization Algorithm (FMH-WOA), in which instead of using conventional whale algorithm, the stochastically parameters are dynamically tuned utilizing a frequency parameter.
Abstract: Multi-robots configuration provides an effective means for space exploration in an obstacle cluttered environment. Recent progression in research involves a significant work in the field of robotics path planning. Advanced algorithms compute these trajectories utilizing two or more techniques. One such approach involves integration of deterministic Coordinated Multi-robot Exploration and meta-heuristic Whale Optimizer, collectively referred as stochastic method/exploration. This research presents an improved version of stochastic exploration namely Frequency Modified Hybrid Whale Optimization Algorithm (FMH-WOA), in which instead of using conventional whale algorithm, the stochastic parameters are dynamically tuned utilizing a frequency parameter. The frequency is adjusted to tune and optimize both the Exploitation and Exploration operators. Deterministic method derives the cost and utility determines the precedence of grid cell around robot. Stochastic optimizer then helps in improving the overall solution. The effectivity of the proposed FMH-WOA is validated under different environmental conditions. The results are then compared with conventional whale optimizer to demonstrate the improvements achieved in terms of enhanced area coverage in considerably less exploration time.
TL;DR: A multifunctional electronic skin (e‐skin) incorporating multiple perceptions with intelligent robotic control is reported, by which robots can safely and dexterously interact with humans.
Abstract: Human–robot collaboration is playing more and more important roles in current deployments of robotic systems in our lives. Haptic perception and intelligent control are essential to ensure safety and efficiency of human–robot interaction. However, existing robotic sensory and control systems are deficient in terms of performance issues, complexity, and cost. Here, the authors report a multifunctional electronic skin (e‐skin) incorporating multiple perceptions with intelligent robotic control, by which robots can safely and dexterously interact with humans. The e‐skin with a simple and cost‐effective sensory structure has multimodal perceptions of proximity, temperature, contact force, and contact position with broad measuring range, high sensitivity, and fast response. The e‐skin is applied onto robots to accomplish obstacle avoidance, safe and dexterous human–robot interaction, smart teaching, and playing Tai‐Chi, which demonstrate a broad range of applications for intelligent robots equipped with e‐skins.
TL;DR: In this paper , a distributed simulation of large-area metasurfaces is proposed to account for scatterer-scatterer interactions, which achieves a linear reduction in the simulation time with the number of compute nodes.
Abstract: Abstract Fast and accurate electromagnetic simulation of large-area metasurfaces remains a major obstacle in automating their design. In this paper, we propose a metasurface simulation distribution strategy which achieves a linear reduction in the simulation time with the number of compute nodes. Combining this distribution strategy with a GPU-based implementation of the Transition-matrix method, we perform accurate simulations and adjoint sensitivity analysis of large-area metasurfaces. We demonstrate ability to perform a distributed simulation of large-area metasurfaces (over 600 λ × 600 λ ), while accurately accounting for scatterer-scatterer interactions significantly beyond the locally periodic approximation.
TL;DR: In this article , the authors used a structured literature review to explore the factors affecting the financial performance of the product take-back system and investigate how 12 factors, clustered into three different dimensions; context, supply chain, and company, affect financial performance.
TL;DR: This study proposed a hybrid multi-target path planning algorithm that was verified by experiments and compared with the other four algorithms in both ordinary and complex environments, demonstrating the strong applicability and high effectiveness of the proposed method.
Abstract: To solve the problem of traversal multi-target path planning for an unmanned cruise ship in an unknown obstacle environment of lakes, this study proposed a hybrid multi-target path planning algorithm. The proposed algorithm can be divided into two parts. First, the multi-target path planning problem was transformed into a traveling salesman problem, and an improved Grey Wolf Optimization (GWO) algorithm was used to calculate the multi-target cruise sequence. The improved GWO algorithm optimized the convergence factor by introducing the Beta function, which can improve the convergence speed of the traditional GWO algorithm. Second, based on the planned target sequence, an improved D* Lite algorithm was used to implement the path planning between every two target points in an unknown obstacle environment. The heuristic function in the D* Lite algorithm was improved to reduce the number of expanded nodes, so the search speed was improved, and the planning path was smoothed. The proposed algorithm was verified by experiments and compared with the other four algorithms in both ordinary and complex environments. The experimental results demonstrated the strong applicability and high effectiveness of the proposed method.
01 Jan 2022
TL;DR: In this article , a target-tracking controller for the UAVs is developed to meet the requirements of payload dynamics and obstacle avoidance, and the authors also present simulation results to demonstrate the benefits of the proposed problem formulation for a multiobstacle environment.
Abstract: Control barrier functions have been widely studied and applied to safety-critical systems, including multi-agent obstacle avoidance problems. In this work, we apply control barrier functions to a collaborative transportation problem involving two unmanned aerial vehicles (UAVs) moving a payload around obstacles as they deliver it to a target location. We develop a target-tracking controller for the UAVs, which is constrained to meet the requirements of payload dynamics and obstacle avoidance. We also present simulation results to demonstrate the benefits of the proposed problem formulation for a multi-obstacle environment.
TL;DR: In this paper , an adaptive barrier Lyapunov function based obstacle avoidance control scheme is proposed for AUV, where the adaptive law is used to approximate the dynamic uncertainties and external disturbances, and the control scheme limits the tracking error within the preset range and thus ensures the safety of the AUV.
TL;DR: It is observed that despite significant progress, deep learning techniques also face difficulties in complex and unknown environments where objects of varying types and shapes are present.
Abstract: Mobile robots lack a driver or a pilot and, thus, should be able to detect obstacles autonomously. This paper reviews various image-based obstacle detection techniques employed by unmanned vehicles such as Unmanned Surface Vehicles (USVs), Unmanned Aerial Vehicles (UAVs), and Micro Aerial Vehicles (MAVs). More than 110 papers from 23 high-impact computer science journals, which were published over the past 20 years, were reviewed. The techniques were divided into monocular and stereo. The former uses a single camera, while the latter makes use of images taken by two synchronised cameras. Monocular obstacle detection methods are discussed in appearance-based, motion-based, depth-based, and expansion-based categories. Monocular obstacle detection approaches have simple, fast, and straightforward computations. Thus, they are more suited for robots like MAVs and compact UAVs, which usually are small and have limited processing power. On the other hand, stereo-based methods use pair(s) of synchronised cameras to generate a real-time 3D map from the surrounding objects to locate the obstacles. Stereo-based approaches have been classified into Inverse Perspective Mapping (IPM)-based and disparity histogram-based methods. Whether aerial or terrestrial, disparity histogram-based methods suffer from common problems: computational complexity, sensitivity to illumination changes, and the need for accurate camera calibration, especially when implemented on small robots. In addition, until recently, both monocular and stereo methods relied on conventional image processing techniques and, thus, did not meet the requirements of real-time applications. Therefore, deep learning networks have been the centre of focus in recent years to develop fast and reliable obstacle detection solutions. However, we observed that despite significant progress, deep learning techniques also face difficulties in complex and unknown environments where objects of varying types and shapes are present. The review suggests that detecting narrow and small, moving obstacles and fast obstacle detection are the most challenging problem to focus on in future studies.