Self-collision Detection Research Articles

Self-Collision Detection and Avoidance for Dual-Arm Concentric Tube Robots

Recent studies on concentric tube robots (CTRs) have shown that they are well-suited for minimally invasive endoscopic surgeries. However, typical surgical procedures require the use of multiple tools simultaneously which has led to the development of dual-arm CTRs that are susceptible to self-collision. In this paper, a closed-loop control system for dual-arm CTRs is proposed to detect and avoid the inter-collision between arms along their entire body. The collision detection module finds the minimum distance between the manipulators in the Cartesian space. To avoid self-collision, the proposed control system using differential Jacobian-based inverse kinematics is developed with three tasks with different priorities; physical constraints, self-collision avoidance, and end-effector tracking. The performance of the proposed control scheme is investigated through implementation for Cartesian control of a dual-arm CTR to reach pre-defined target points in a simulated scenario similar to epilepsy surgery. The self-collision detection module successfully predicted all 359 self-collision cases in the target region. The experimental results demonstrated the efficacy of the controller in handling inter-collision between arms over their entire body by keeping the minimum distance between arms at 1.542 mm.

Neural Network Method of Controlling Self-Collisions of Multilink Manipulators Based on the Solution of the Classification Problem

The problem of self-collisions of manipulators with several links installed on a robot can arise when they work together in one zone. To prevent self-collisions, it is necessary to develop methods for their detection and their subsequent inclusion in control algorithms. This paper proposes an approach for determining the occurrence of self-collisions of manipulators using the Artificial Neural Networks approach. In contrast to the regression problem, this article proposes a classification approach. The effectiveness of the proposed approach was tested on robots with multilink manipulators “Ichtiandr” and SAR-401 and their simulators. Self-collision detection using the proposed method is much faster than using the traditional approach of solving the inverse kinematics problem without loss of accuracy. The problem was solved by constructing various Artificial Neural Networks and then checking the accuracy of the solution. A comparative analysis of Artificial Neural Networks was carried out and as a result, the Artificial Neural Networks approach showing the best accuracy was selected. The problem was solved for a robot with two manipulators. The resulting solution can be extended to a larger number of manipulators installed on the robot.

A Self-Collision Detection Algorithm of a Dual-Manipulator System Based on GJK and Deep Learning.

Self-collision detection is fundamental to the safe operation of multi-manipulator systems, especially when cooperating in highly dynamic working environments. Existing methods still face the problem that detection efficiency and accuracy cannot be achieved at the same time. In this paper, we introduce artificial intelligence technology into the control system. Based on the Gilbert-Johnson-Keerthi (GJK) algorithm, we generated a dataset and trained a deep neural network (DLNet) to improve the detection efficiency. By combining DLNet and the GJK algorithm, we propose a two-level self-collision detection algorithm (DLGJK algorithm) to solve real-time self-collision detection problems in a dual-manipulator system with fast-continuous and high-precision properties. First, the proposed algorithm uses DLNet to determine whether the current working state of the system has a risk of self-collision; since most of the working states in a system workspace do not have a self-collision risk, DLNet can effectively reduce the number of unnecessary detections and improve the detection efficiency. Then, for the working states with a risk of self-collision, we modeled precise colliders and applied the GJK algorithm for fine self-collision detection, which achieved detection accuracy. The experimental results showed that compared to that with the global use of the GJK algorithm for self-collision detection, the DLGJK algorithm can reduce the time expectation of a single detection in a system workspace by 97.7%. In the path planning of the manipulators, it could effectively reduce the number of unnecessary detections, improve the detection efficiency, and reduce system overhead. The proposed algorithm also has good scalability for a multi-manipulator system that can be split into dual-manipulator systems.

An effective self-collision detection algorithm for multi-degree-of-freedom manipulator

In order to address the self-collision problem associated with the operation of modern industrial robots, this paper proposes a multi-degree-of-freedom collision detection algorithm that can detect self-collision in single arm and double arm robots as well as collision with the load. Firstly, the zero pose of the Denaviti−Hartenberg model is built based on the manipulator configuration, and the coordinate information of each key point is obtained through a rotation and translation operation of the matrix. Then, the positional relation and distance between the detected objects are determined by the spatial geometry theory, and finally, collision is detected using a collision matrix. By simulating two groups of single arms and two groups of double arms, and from the laboratory testing of SR10C in the SIASUN robot factory, it has been verified that the proposed algorithm has good collision detection capability. Without the use of sensors, cameras, and other external devices, the collision between the arm and the load, and the collision between the cooperative robot and the load may be effectively detected and mitigated.

Optimizing Kinematic Modeling and Self-Collision Detection of a Mobile Manipulator Robot by Considering the Actual Physical Structure

In this paper, an optimized kinematic modeling method to accurately describe the actual structure of a mobile manipulator robot with a manipulator similar to the universal robot (UR5) is developed, and an improved self-collision detection technology realized for improving the description accuracy of each component and reducing the time required for approximating the whole robot is introduced. As the primary foundation for trajectory tracking and automatic navigation, the kinematic modeling technology of the mobile manipulator has been the subject of much interest and research for many years. However, the kinematic model established by various methods is different from the actual physical model due to the fact that researchers have mainly focused on the relationship between driving joints and the end positions while ignoring the physical structure. To improve the accuracy of the kinematic model, we present a kinematic modeling method with the addition of key points and coordinate systems to some components that failed to model the physical structure based on the classical method. Moreover, self-collision detection is also a primary problem for successfully completing the specified task of the mobile manipulator. In traditional self-collision detection technology, the description of each approximation is determined by the spatial transformation of each corresponding component in the mobile manipulator robot. Unlike the traditional technology, each approximation in the paper is directly established by the physical structure used in the kinematic modeling method, which significantly reduces the complicated analysis and shortens the required time. The numerical simulations prove that the kinematic model with the addition of key point technology is similar to the actual structure of mobile manipulator robots, and the self-collision detection technology proposed in the article effectively improves the performance of self-collision detection. Additionally, the experimental results prove that the kinematic modeling method and self-collision detection technology outlined in this paper can optimize the inverse kinematics solution.

Self-collision detection algorithm based on fused DNN and AABB-circular bounding box

Self-collision detection algorithm based on fused DNN and AABB-circular bounding box

Dynamics modeling and simulation of self-collision of tether-net for space debris removal

Tether-net is a new active removal technology for space debris, and its deployment and capture have attracted considerable attention. This study focuses on the dynamics and simulation of self-collision of tether-net. First, the mass-spring-damper method is used to model tether-net and a line–line self-collision detection algorithm is proposed according to the geometric characteristics of tether-net. Thereafter, combined with the nonlinear collision model, the self-collision process of tether-net is studied. Two simulations of the close-up of a net with or without a target are executed to show the difference between considering and not considering the self-collision of tether-net. Results reveal that the capture process of tether-net with consideration for self-collision is different from the one without self-collision, especially after the corners of the net begin to contact each other.

Kinematics analysis and self - collision detection of Truss type multi-robot cooperative welding platform

A collaborative welding platform consists of a XYZ-mobile truss, a 3R (revolute joint) clamping robot and a dual-arm welding robot. Based on Denavit-Hartenberg (D-H) method, the analytical inverse kinematics solution of the platform is obtained. Dual-arm coordination is a technical difficulty in robot control field. Focusing on self-collision problem during dual-arm robot motion planning, this paper proposes a new obstacle avoidance algorithm based on Rapidly Exploring Random Tree (RRT) algorithm. Setting searching area founded on artificial potential field (APF) algorithm and adopting note-pruning strategy achieves high searching efficiency. Simulation results verify the correctness of kinematics analysis and self-collision detection algorithm.

DeepCrash: A Deep Learning-Based Internet of Vehicles System for Head-On and Single-Vehicle Accident Detection With Emergency Notification

Most individuals involved in traffic accidents receive assistance from drivers, passengers, or other people. However, when a traffic accident occurs in a sparsely populated area or the driver is the only person in the vehicle and the crash results in loss of consciousness, no one will be available to send a distress message to the proper authorities within the golden window for medical treatment. Considering these issues, a method for detecting high-speed head-on and single-vehicle collisions, analyzing the situation, and raising an alarm is needed. To address such issues, this paper proposes a deep learning-based Internet of Vehicles (IoV) system called DeepCrash, which includes an in-vehicle infotainment (IVI) telematics platform with a vehicle self-collision detection sensor and a front camera, a cloud-based deep learning server, and a cloud-based management platform. When a head-on or single-vehicle collision is detected, accident detection information is uploaded to the cloud-based database server for self-collision vehicle accident recognition, and a related emergency notification is provided. The experimental results show that the accuracy of traffic collision detection can reach 96% and that the average response time for emergency-related announcements is approximately 7 s.

Accurate self-collision detection using enhanced dual-cone method

We present an accurate and robust algorithm for self-collision detection in deformable models. Our method is based on the normal cone test and is suitable for both discrete and continuous collision queries on triangular meshes. We propose a novel means of employing surface normal cones and binormal cones to perform the normal cone test. Moreover, we combine our culling criteria with bounding volume hierarchies (BVHs) and present a hierarchical traversal scheme. Unlike the previous BVH-based dual-cone method, our method can reliably detect all self-collisions, and it achieves appreciable speedup over other high-level culling methods.

A New Self-Collision Detection Method for Cloth Simulation

In the system of cloth simulation, the process of cloth deformation and the method of self-collision detection have great significance for fast cloth deformation and realistic rendering. In this paper, we first presented a novel contour test based on the star-shaped property of a polygon, i.e. the central data structure for efficient execution of the star-contour test. Then, the paper describes and compares in detail, directly accumulation method and global stiffness matrix method. Taking efficiency, stability and convergence into consideration, the paper puts forward an explicit-implicit combined method which is suitable for force models. At last, a fabric deformation constraint method has been designed, which is used for linear FEM. This method is effective to restrain the fabric from excessive deformation behavior and create realistic simulation results.

GPU-based radial view-based culling for continuous self-collision detection of deformable surfaces

We propose a graphics processing unit-based approach to accelerate the radial view-based culling method for continuous self-collision detection of deformable surfaces. The deformable surfaces may have small round-shaped holes and ghost triangles are used to fill the holes. We identify the key processes of the radial view-based culling method, including triangle classification, traversal of bounding volume hierarchies and handling violated triangles (i.e., the triangles intersecting with ghost triangles). We propose efficient parallel processing techniques to perform these key processes on a programmable graphics unit. We have evaluated our proposed approach on several examples. Experimental results show that our approach significantly cuts down the cost of the key processes of the radial-based culling method, compared with the serial implementation on CPU.

An Optimized Fabric Self Collision Detection Method

In order to deal with self-collision detection in fabric dynamic simulation, we provide a set of optimized strategies. In rough detection phase, we use the continuous normal vector cone method applicable to quad bounding volume tree. The method uses the normal vector cone information between every two sub-trees for the pruning, which can filter out more areas where intersection is impossible, thus can improve the efficiency of collision detection. In accurate detection phase, we provide a set of reasonable detection strategies for the primitives within the time step, which can enhance the efficiency of while effectively ensure the correctness of collision detection. Our method can effectively deal with self-collision in the dynamic simulation with the good sense of reality.

1D sweep-and-prune self-collision detection for deforming cables

Detecting self-collision for cables and similar objects is an important part of numerous models in computational biology (protein chains), robotics (electric cables), hair modeling, computer graphics, etc. In this paper the 1D sweep-and-prune algorithm for detecting self-collisions of a deforming cable comprising linear segments is investigated. The sweep-and-prune algorithm is compared with other state-of-the-art self-collision detection algorithms for deforming cables and is shown to be up to an order of magnitude faster than existing algorithms for cables with a high proportion of segments moving. We also present a multi-threaded version of the algorithm and investigate its performance. In addition, we present worst-case bounds for 1D sweep-and-prune algorithms whereby the colliding objects do not exceed a certain object density, and apply these results to deforming cables.

Minimum Normal Cone for Continuous Collision Detection

Continuous collision detection (CCD) algorithm is a vital step in virtual reality. This paper describes an efficient self-collision detection algorithm called the minimum normal cone algorithm. The algorithm calculates the parent-normal cone in a new way, which optimizes the two test conditions for the self-collision detection: In the first test condition, we analyze and calculate the relative position between the two child-normal cones, and then a bottom-up technology is used to build the parent-normal cone. Thus the two child normal cones can be surrounded by the parent-normal cone perfectly, and the normal vectors motion range of the parent-triangular mesh can also be closely restricted; In the second test condition, we calculate the normal vectors motion range of the contour projection line, and check whether the projection line is self-colliding or not, to avoid excessive tests between segments. The algorithm has been implemented and tested on two classical models, cloth-ball and N-body model respectively. Experimental results demonstrate that our algorithm can decrease the number of elementary tests by two orders of magnitudes, and significantly improve the performance of the overall CCD algorithm.

Radial view based culling for continuous self-collision detection of skeletal models

We present a novel radial-view-based culling method for continuous self-collision detection (CSCD) of skeletal models. Our method targets closed triangular meshes used to represent the surface of a model. It can be easily integrated with bounding volume hierarchies (BVHs) and used as the first stage for culling non-colliding triangle pairs. A mesh is decomposed into clusters with respect to a set of observer primitives (i.e., observer points and line segments) on the skeleton of the mesh so that each cluster is associated with an observer primitive. One BVH is then built for each cluster. At the runtime stage, a radial view test is performed from the observer primitive of each cluster to check its collision state. Every pair of clusters is also checked for collisions. We evaluated our method on various models and compared its performance with prior methods. Experimental results show that our method reduces the number of the bounding volume overlapping tests and the number of potentially colliding triangle pairs, thereby improving the overall process of CSCD.

Energy-based self-collision culling for arbitrary mesh deformations

In this paper, we accelerate self-collision detection (SCD) for a deforming triangle mesh by exploiting the idea that a mesh cannot self collide unless it deforms enough. Unlike prior work on subspace self-collision culling which is restricted to low-rank deformation subspaces, our energy-based approach supports arbitrary mesh deformations while still being fast. Given a bounding volume hierarchy (BVH) for a triangle mesh, we precompute Energy-based Self-Collision Culling (ESCC) certificates on bounding-volume-related sub-meshes which indicate the amount of deformation energy required for it to self collide. After updating energy values at runtime, many bounding-volume self-collision queries can be culled using the ESCC certificates. We propose an affine-frame Laplacian-based energy definition which sports a highly optimized certificate pre-process, and fast runtime energy evaluation. The latter is performed hierarchically to amortize Laplacian energy and affine-frame estimation computations. ESCC supports both discrete and continuous SCD with detailed and nonsmooth geometry. We observe significant culling on many examples, with SCD speed-ups up to 26X.

Energy-based self-collision culling for arbitrary mesh deformations

In this paper, we accelerate self-collision detection (SCD) for a deforming triangle mesh by exploiting the idea that a mesh cannot self collide unless it deforms enough. Unlike prior work on subspace self-collision culling which is restricted to low-rank deformation subspaces, our energy-based approach supports arbitrary mesh deformations while still being fast. Given a bounding volume hierarchy (BVH) for a triangle mesh, we precompute Energy-based Self-Collision Culling (ESCC) certificates on bounding-volume-related sub-meshes which indicate the amount of deformation energy required for it to self collide. After updating energy values at runtime, many bounding-volume self-collision queries can be culled using the ESCC certificates. We propose an affine-frame Laplacian-based energy definition which sports a highly optimized certificate pre-process, and fast runtime energy evaluation. The latter is performed hierarchically to amortize Laplacian energy and affine-frame estimation computations. ESCC supports both discrete and continuous SCD with detailed and nonsmooth geometry. We observe significant culling on many examples, with SCD speed-ups up to 26X.

Dynamic self-collision detection and prevention for 2-DOF robot arms using interval-based analysis

The problems related to self-collision detection and optimal collision-free trajectory planning for a robot arm subjected to dynamic constraints is investigated. First, a computed-torque method is used to obtain a linearized closed-loop system. For this linearized system, the reference state that the robot arm is capable of reaching is verified through phase plane analysis. This will ensure that the robot arm can be stopped before self-collision occurs. Dynamic constraints are taken into account for a continuous motion of deceleration by calculating the bounds of the commanded force/torques with interval evaluations. When the reference state e at t+δt is not valid for selfcollision avoidance, a new feasible state is determined by adhering to an interval-based method which allows decomposition of a complex constrained optimization problem into a simple two-stage optimization problem with relaxed constraints. The optimized feasible state not only secures the robot arm against self-collision but also allows the robot arm to track the original reference trajectory closely. Simulation and experimental results of a 2-dof robot arm show the effectiveness of the proposed interval-based approach.

Real-time cloth simulation based on particle constraints

A physics-based cloth simulation using modified mass-spring model was presented.There was no force generated by springs between particles,but a set of constraints was built between particles.And real-time cloth simulation was achieved according to physical rules to continually adjust the positions of the particles to satisfy these constraints.Based on the spatial hash function,the cloth self-collision detection and response were solved.Using NVIDIA video's CUDA technology,hardware acceleration on cloth simulation was implemented and its frame rate was increased about several times.