Collision Avoidance Research Articles

Stable Biped Robot's Walk using Semi-Supervised ANN based Trajectory Generation within Yolov5 Algorithm based Identified Environment with Ditch

The study introduces a stable walking pattern for a biped robot by employing a semi-supervised artificial neural network (ANN) to generate trajectories with a focus on reducing potential damage from small objects that are identified by Yolov5 algorithm. The ANN is utilized as a universal approximator to ensure smooth motion automatically by meeting predefined boundary conditions during its training. This trajectory generation approach is then compared with one another ANN- based method, with nonstop evaluations mainly focusing on position, velocity, and its acceleration profiles to maintain smooth motion. By analysis of trajectory derivatives and its curvature detects and auto corrects any discontinuities. Mathematical model created on from MATLAB 2023 and its simulations validate the trajectory's smoothness and demonstrating its effectiveness in enabling bipedal robots to navigate uneven terrain. The proposed method is very useful and more suitable for online adaptable trajectory generation by addressing collision avoidance and adaptability to various terrains, and overall stability in bipedal robot navigation comprehensively.

Establishing requirements for lunar and cislunar orbital debris tracking

There is a need to formally coordinate lunar and cislunar space traffic management to mitigate the risk of collisions with micrometeoroids and orbital debris in this space domain. To control this hazard, it will be critical to develop a high-fidelity orbital debris catalog. This catalog will be maintained by monitoring and propagating the trajectory of objects. The need to perform a debris avoidance maneuver for collision avoidance will depend on the fidelity of the propagated debris trajectory. A larger uncertainty (magnitude as a function of time) will require a larger maneuver and a higher likelihood of the need to perform this maneuver. This study assesses debris avoidance maneuvers and corresponding corrective actions to recover the desired mission trajectory as a way to evaluate the desired capability of tracking of objects in lunar and cislunar orbit. The baseline Earth-to-Moon trajectory was that of the Artemis I mission. Typical conjunction assessments were postulated at 100,000 km from Earth's center (approximately a quarter of the way to the moon) and at lunar orbit insertion (LOI). For these two cases, the required debris avoidance maneuver (and hence impact to the mission) is tied to the uncertainty in orbital debris tracking. The study provides a methodology and baseline inputs to establish future requirements of debris tracking in lunar and cislunar orbits. It also reinforces the importance of long-term sustainability for lunar missions, specifically preventing the generation of orbital debris.

Distributed fixed-time control for high-order multi-agent systems with FTESO and feasibility constraints

In this study, a fixed-time distributed anti-disturbance control strategy is designed for a class of higher-order multi-agent systems in an undirected static topology. Inspired by the existing literature, the strategy introduces a unique way of defining neighborhood errors during the design process to deal with the connectivity maintenance and collision avoidance problems among the neighbor agents in the system. The control strategy not only solves the full state constraint and transient steady state performance constraint control problems simultaneously, but also considers the effects of internal uncertainties and external disturbances of the system, which enhances the difficulty of controller design. To address these challenges, a fixed-time extended state observer is introduced to obtain the estimation of uncertainty and disturbance . Subsequently, an integrated backstepping control scheme is designed to feed-forward compensate for uncertainties and disturbances, while constraining the transient steady-state performance of the system to remain within a predefined performance boundaries. This is achieved by combining the fixed-time prescribed performance and the barrier Lyapunov function. Finally, a simulation case is presented to verify the effectiveness of the designed control strategy.

AISClean: AIS data-driven vessel trajectory reconstruction under uncertain conditions

In maritime transportation, intelligent vessel surveillance has become increasingly prevalent and widespread by collecting and analyzing high massive spatial data from automatic identification system (AIS). The state-of-the-art AIS devices contain various functionalities, such as position transmission, tracking navigation, etc. Widely equipped shipboard AIS devices provide a large amount of real-time and historical vessel trajectory data for maritime management. However, the original AIS data often suffers from unwanted noise (i.e., poorly tracked timestamped points for vessel trajectories) and missing (i.e., no data is received or transmitted for a long term) data during signal acquisition, transmission, and analog-to-digital conversion. This degradation in data quality poses significant risks, including potential miscalculations in vessel collision avoidance systems, inaccuracies in emission calculations, and challenges in port management. In this work, a data-driven vessel trajectory reconstruction framework considering historical features is proposed to enhance the reliability of vessel trajectory. Specifically, a series of statistical methods are proposed to identify noisy data and missing data. Then, a model combining Geohash and dynamic time warping algorithms is developed to restore the trajectories degraded by random noise and missing data in vessel trajectories. Comparative experiments with baseline methods on multiple datasets verify the effectiveness of the proposed data-driven model.

Non-monotonic behavior of jam probability and stretched exponential distribution in pedestrian counterflow

Abstract We adopt a floor field cellular automata model to study the statistical properties of bidirectional pedestrian flow moving in a straight corridor. We introduce a game-theoretic framework to deal with the conflict of multiple pedestrians trying to move to the same target location. By means of computer simulations, we show that the complementary cumulative distribution of the time interval between two consecutive pedestrians leaving the corridor can be fitted by a stretched exponential distribution, and surprisingly, the statistical properties of the two types of pedestrian flows are affected differently by the flow ratio, i.e., the ratio of the pedestrians walking toward different directions. We also find that the jam probability exhibits a non-monotonic behavior with the flow ratio, where the worst performance arises at an intermediate flow ratio of around 0.2. Our simulation results are consistent with some empirical observations, which suggest that the peculiar characteristics of the pedestrians may attributed to the anticipation mechanism of collision avoidance.

Identification of Complex Multi-Vessel Encounter Scenarios and Collision Avoidance Decision Modeling for MASSs

Complex multi-vessel encounter situations are a challenging problem for ships to avoid collisions, and the International Regulations for Preventing Collision at Sea, 1972 (COLREGs) do not provide a clear delineation of multi-vessel encounter situations and the responsibility of collision avoidance (CA). Furthermore, Marine Autonomous Surface Ships (MASS), which realize autonomous navigation functions, face the problem of recognizing complex multi-ship encounter situations and the corresponding CA decisions. In this study, we adopt the velocity obstacle (VO) algorithm to visualize and identify the danger of multi-ship encounters with the own ship (OS) as the first viewpoint. Additionally, we consider the motion changes in target ships (TSs) and their possible CA behaviors as the basis of the ship’s CA decision-making. According to COLREGs, a simplified method for classifying the encounter situations of multiple clustered ships is proposed, considering the coupling of collision hazards and CA responsibilities between related TSs. On this basis, the corresponding CA decisions for each classified situation are proposed, and a large number of simulation experiments are conducted based on the proposed method by considering the three-ship and four-ship encounter model in the Imazu problem as an example. The experimental results indicate that the proposed method can effectively recognize the complex multi-ship encounter situation in the Imazu problem, and it can adjust the CA measures of the OS in time according to the COLREGs and the behavior of TSs. This provides the basis and reference for MASS when facing complex multi-ship encounter situations.

LiDAR-Based Urban Three-Dimensional Rail Area Extraction for Improved Train Collision Warnings.

The intrusion of objects into track areas is a significant issue affecting the safety of urban rail transit systems. In recent years, obstacle detection technology based on LiDAR has been developed to identify potential issues, in which accurately extracting the track area is critical for segmentation and collision avoidance. However, because of the sparsity limitations inherent in LiDAR data, existing methods can only segment track regions over short distances, which are often insufficient given the speed and braking distance of urban rail trains. As such, a new approach is developed in this study to indirectly extract track areas by detecting references parallel to the rails (e.g., tunnel walls, protective walls, and sound barriers). Reference point selection and curve fitting are then applied to generate a reference curve on either side of the track. A centerline is then extrapolated from the two curves and expanded to produce a 2D track area with the given size specifications. Finally, the 3D track area is acquired by detecting the ground and removing points that are either too high or too low. The proposed technique was evaluated using a variety of scenes, including tunnels, elevated sections, and level urban rail transit lines. The results showed this method could successfully extract track regions from LiDAR data over significantly longer distances than conventional algorithms.

Neural network-based control for the on-orbit assembly of heterogeneous spacecraft cluster based on Vicsek fractal

On-orbit assembly is a highly effective technique for building large space structures. This paper presents the dynamics and control of an on-orbit assembly of a large space structure based on a heterogenous spacecraft cluster comprising both rigid and flexible spacecraft. The topology of the large space structure is inspired by the Vicsek fractal. A distributed assembly strategy is proposed for the spacecraft team. Radial Basis Function neural networks (RBFNNs) are utilized to approximate the bounded uncertain terms in translational and attitude dynamics. To avoid collisions during the pre-assembly phase, a neural network (NN)-based controller with collision avoidance force is designed for relative position motion. During the assembly phase, only proportional-derivative (PD) law is employed for relative position control. In addition, an NN-based controller is designed for relative attitude control for both rigid and flexible spacecraft. To estimate the unmeasured flexible vibrations, modal coordinate observers are introduced for the flexible spacecraft. The stability of the closed-loop system is proved via Lyapunov functions. Moreover, numerical results are presented to validate the effectiveness of the proposed controllers and assembly strategy.

Bidirectional Long Short-Term Memory Development for Aircraft Trajectory Prediction Applications to the UAS-S4 Ehécatl

The rapid advancement of unmanned aerial systems in various civilian roles necessitates improved safety measures during their operation. A key aspect of enhancing safety is effective collision avoidance, which is based on conflict detection and is greatly aided by accurate trajectory prediction. This paper represents a novel data-driven trajectory prediction methodology based on applying the Long Short-Term Memory (LSTM) prediction algorithm to the UAS-S4 Ehécatl. An LSTM model was designed as the baseline and then developed into a Staked LSTM to better capture complex and hierarchical temporal trajectory patterns. Next, the Bidirectional LSTM was developed for a better understanding of the contextual trajectories from both its past and future data points, and to provide a more comprehensive temporal perspective that could enhance its accuracy. LSTM-based models were evaluated in terms of mean absolute percentage errors. The results reveal the superiority of the Bidirectional LSTM, as it could predict UAS-S4 trajectories more accurately than the Stacked LSTM. Moreover, the developed Bidirectional LSTM was compared with other state-of-the-art deep neural networks aimed at aircraft trajectory prediction. Promising results confirmed that Bidirectional LSTM exhibits the most stable MAPE across all prediction horizons.

Collision Avoidance of UAV Formation Using Geometric Approach and Relative Velocity Vector in Three-dimensional Space

Collision Avoidance of UAV Formation Using Geometric Approach and Relative Velocity Vector in Three-dimensional Space

Internet of robotic things with a local LoRa network for teleoperation of an agricultural mobile robot using a digital shadow

In unstructured agricultural fields where autonomous navigation is challenging and demands additional safety, the operator’s experience and knowledge are essential for supervising operations and making decisions beyond the robot’s autonomous capabilities. Local networks with long-range wireless communication combined with digital twin concepts are promising solutions that can be used for robot teleoperation. The purpose of this study was to demonstrate the feasibility of supervising a mobile robot inside berry orchards using a digital shadow from a long-range distance (between 300 and 3000 m), with the primary objective of assisting the robot in navigating in complex situations such as row-end turning. This involved creating a virtual representation of the robot that mirrors its state and actions, allowing the remote operator to monitor and guide the robot effectively. The system comprised a GPS-based navigation controller with collision avoidance sensors, two sets of LoRa transmitters and repeaters, a simulation environment with a digital shadow of the robot, and a graphical user interface for the remote operator. Information about the digital shadow’s state, including location, orientation, and distances to obstacles, was received as a message by the LoRa gateway and was used to update the path for the actual robot that interfaced with the Robot Operating System (ROS). The main research hypothesis aimed to test the quality of the LoRa communication link between the robot and the operator, as well as the robustness of the robot’s control system, with an emphasis on the architecture, communication link, and situation awareness creation. Preliminary results showed that depending on the environment, the average packet loss was 12% at distances of approximately 2300 m. Our results highlight some of the core technical challenges that need to be addressed for an effective teleoperation system, including latency, stability, and the limited range of wireless communication. Future works involves evaluating the performance and reliability of the proposed method under different field conditions and scenarios, as well as considering the use of the 5G network for a significant improvement in data transmission speed, navigation efficiency, and visual feedback. Upon successful implementation, this study has the potential to enhance the efficiency and safety of robot navigation, providing a practical solution for remote supervision in challenging environments.

USV formation navigation decision-making through hybrid deep reinforcement learning using self-attention mechanism

To address the challenging of balancing Unmanned Surface Vessel (USV) autonomous collision avoidance and formation maintenance in uncertain environments, a formation construction and navigation decision-making strategy based on Hybrid Deep Reinforcement Learning (HDRL) is proposed in this study. The novelty of this study is that: (1) A HDRL training approach is proposed, incorporating diverse DRL for virtual leader and followers, thereby significantly enhancing the decision-making adaptability of USVs formation. (2) The multi-head attention mechanism of decentralized Critic strategy is used to enhance the attentional focus of the HDRL algorithm towards different agents, thereby effectively improving convergence speed. (3) The method employs a meticulously designed hybrid reward function and incorporates the Optimal Reciprocal Collision Avoidance (ORCA) for speed selection, thereby providing optimal speed recommendations based on the current situation. It is worth emphasizing that the method exhibits exceptional performance in terms of success rate, navigation time, average reward value, and other relevant indicators within the simulation. Finally, the method is validated through the utilization of real-time navigation data in the Panama Canal, thereby substantiating the potential engineering applicability.

Long-Term Fuel-Optimal Collision Avoidance Maneuvers with Station-Keeping Constraints

This work presents a sequential convex program method to compute fuel-optimal collision avoidance maneuvers for long-term encounters. The low-thrust acceleration model is used to account for the control, but the method can compute high-thrust maneuvers by increasing the maximum available acceleration. Dealing with the long-term conjunction poses additional challenges compared to the short-term problem because the encounter is not instantaneous. Thus, under the assumption of Gaussian statistics, the probability of collision is replaced by a simpler metric, the instantaneous probability of collision (IPoC), and a keep-out zone constraint is formulated as a continuous condition to be respected throughout the time frame of interest. The robustness of the solution is improved by introducing a constraint on the sensitivity of IPoC. Furthermore, the collision avoidance problem is coupled with the classical station-keeping requirement for geostationary Earth orbit satellites and with a return to the nominal orbit condition for low Earth orbit satellites. Even though no guarantee is given for the recovery of the global optimum solution, numerical simulations in different orbital regimes show that the proposed approach can yield a local fuel-optimal solution with a run time suitable for autonomous applications.

A behavioral conditional diffusion probabilistic model for human motion modeling in multi-action mixed human-robot collaboration

Collision avoidance is priority for human-robot collaboration (HRC). Human motions have stochastic, making it difficult for robots to recognize humans. To establish accurate human motion models is important to recognize humans. In the collaboration process, the action categories may suddenly and continuously switch to adapt to complex production tasks, namely, the stochastic of human motions, and we call this collaboration process as multi-action mixed HRC. The previous HRC rarely considered the stochastic of human motions. However, the stochastic in multi-action mixed HRC brings great challenges to human motion modeling. On the one hand, the same observed motions may develop into different action categories, the motion distribution need to be predicted. On the other hand, the connected motions between two action categories are missing, and need to be predicted. To address these two problems, this study introduced diffusion probabilistic models to capture the stochastic of human motions. However, the current diffusion probabilistic models cannot realize the high-accuracy human motion prediction. Most of them did not consider the condition distribution in diffusion process, and did not comprehensively capture the spatiotemporal characteristics between human and robot. To solve the above problems, a behavioral conditional probabilistic diffusion model (BCDPM) is proposed. The BCDPM model can capture the stochastic of human motions to model the human motion distribution. A mask mechanism is combined with diffusion prior generated by BCDPM model to predict the missing connected motions. The BCDPM model consists of the designed constrain network, the denoising network, and enhanced memory graph neural network. The constrain network can generate the condition distribution, the denoising network and enhanced memory graph neural network can capture the short term and long term spatiotemporal characteristics, respectively. Moreover, the overlapping motion primitives are explicitly encoded into the implicit BCDPM model to mine commonalities of motions. The results show the proposed method can obtain excellent accuracy in prediction and outstanding F1 score in collision detection experiment.

MPC-based motion control of AGV with improved A* and artificial potential field

Nowadays, automatic guided vehicles (AGV) are extensively utilized for transportation and inspection tasks in workshops. The A* and artificial potential field (APF) are classic algorithms employed for path planning of AGVs. However, these algorithms still fail to meet the actual production needs and cannot avoid stuttering while encountering obstacles, leading to excessive energy consumption and unnecessary pause. In the paper, an improved A* algorithm is proposed to reduce route length and improving efficiency. On this basis, an integrated fusion strategy consisting of improved APF and nonlinear model predictive control (NMPC) is designed for collision avoidance and path tracking control. The proposed algorithm is tested both on simulation and a laser-guided real automatic guided vehicle experimental platform. Experimental results prove that the proposed algorithm has a great tracking performance under complex workplace.

Looming Detection in Complex Dynamic Visual Scenes by Interneuronal Coordination of Motion and Feature Pathways

Detecting looming signals for collision avoidance encounters challenges in real‐world scenarios, where moving backgrounds can interfere as an agent navigates through complex natural environments. Remarkably, even insects with limited neural systems adeptly respond to looming stimuli while in motion at high speeds. Existing insect‐inspired looming detection models typically rely on either motion‐pathway or feature‐pathway signals, yet both are susceptible to dynamic visual scene interference. Coordinating interneuron signals from both pathways can enhance the looming detection performance under dynamic conditions. An artificial neural network is employed to construct a combined‐pathway model based on Drosophila anatomy. The model outperforms state‐of‐the‐art bio‐inspired looming‐detection models in tasks involving dynamic backgrounds, simulated by animated 2D‐moving natural scenes or recorded in reality when an unmanned aerial vehicle performs obstacle collision avoidance tasks. Notably, by combining neural anatomy architecture and appropriate multiobjective tasks, the model exhibits convergent neural dynamics with biological counterparts post‐training, offering network explanations and mechanistic insights. Specifically, a multiplicative interneuron operation enhances looming signal patterns and reduces background interferences, generalizing to more complex scenarios, such as AirSim 3D environments and real‐world situations. The work introduces testable biological hypotheses and a promising bioinspired solution for looming detection in dynamic visual environments.

Leader-Following Connectivity Preservation and Collision Avoidance Control for Multiple Spacecraft with Bounded Actuation

This paper investigates the distributed formation control of a group of leader-following spacecraft with bounded actuation and limited communication ranges. In particular, connectivity-preserving and collision-avoidance controllers are proposed for the leader with constant or time-varying velocity, respectively. The communication graph between the spacecraft is modeled via a distance-induced proximity graph. By designing a virtual proxy for each spacecraft, the spacecraft–proxy couplings address the actuator saturation constraints. The inter-proxy dynamics incorporated with a bounded artificial potential function fulfill the coordination of all proxies. In addition, the bounded potential function can simultaneously tackle connectivity preservation and collision avoidance problems. The distributed formation controllers are proposed for multiple spacecraft with constant or time-varying velocities relative to the leader. A sliding mode control approach and the proxies’ dynamics are used in the design of a distributed cooperative controller for spacecraft to address the cooperative problem between the followers and the leader. Numerical simulations confirm the effectiveness of the anti-saturation distributed connectivity preservation controller.

Ship autonomous collision avoidance decision from the perspective of navigation practice

Navigation safety is an eternal theme for ships, and it is an inherent strategic requirement for both transportation powers and maritime powers. With the development of intelligent ships, how to make intelligent decision-making results consistent with maritime practice is the key to navigation safety. This paper constructs a complex water area collision avoidance decision-making model that conforms to maritime practice, providing autonomous collision avoidance decision-making schemes for manned or unmanned ships. Firstly, based on the nonlinear velocity obstacle method, dynamic constraints such as ship maneuverability, target ship motion uncertainty, Convention on the International Regulations for Preventing Collisions at Sea (COLREGs), and good seamanship are fully considered, and a safe collision avoidance action plan is solved in the velocity domain; Then, by introducing an overshoot coefficient to improve the line of sight method, a feasible collision avoidance decision plan is derived by searching for a resumption navigation plan based on the safe avoidance action plan; Finally, collision avoidance decision plan is optimized to achieve autonomous collision avoidance decision-making for ships. To verify the effectiveness of the proposed model, a series of case studies, which focus on different encounter situations are designed and executed, the results indicate that the outcome of the model consistently aligns with real-world scenarios, which is in line with maritime practice.

Hybrid dynamical systems logic and its refinements

Hybrid dynamical systems describe the mixed discrete dynamics and continuous dynamics of cyber-physical systems such as aircraft, cars, trains, and robots. To justify correctness properties of the safety-critical control algorithms for their physical models, differential dynamic logic (▪) provides deductive specification and verification techniques implemented in the theorem prover ▪. The logic ▪ is useful for proving, e.g., that all runs of a hybrid dynamical system α satisfy safety property φ (i.e., ▪), or that there is a run of the hybrid dynamical system α ultimately reaching the desired goal φ (i.e., ▪). Logical combinations of ▪'s operators naturally represent safety, liveness, stability and other properties. Variations of ▪ serve additional purposes. Differential refinement logic (▪) adds an operator α≤β expressing that hybrid system α refines hybrid system β, which is useful, e.g., for relating concrete system implementations α to their abstract verification models β. Just like ▪, ▪ is a logic closed under all operators, which opens up systematic ways of simultaneously relating systems and their properties, of reducing system properties to system relations or, vice versa, reducing system relations to system properties. A second variant of ▪, differential game logic (▪), adds the ability of referring to winning strategies of players in hybrid games, which is useful for establishing correctness properties where the actions of different agents may interfere either because they literally compete with one another or because they may interact accidentally. In the theorem prover ▪, ▪ and its variations have been used for verifying ground robot obstacle avoidance, the Federal Aviation Administration's Next-Generation Airborne Collision Avoidance System ACAS X, and the Federal Railroad Administration's train control model.

Event-triggered based adjustable containment control scheme design for multi-agent systems

The traditional containment control problem uses a mapping method with a fixed communication topology to generate a set of reference signals for followers, which may lead to a situation where the position of the reference signal cannot be adjusted and causes collisions. Therefore, based on backstepping theory, we propose a novel containment control scheme to address this problem, where the generated reference signals can be adjusted arbitrarily. In particular, it is important to point out that an estimator is designed to guide the follower if the reference signal is not available, so that the collision problem is avoided. In addition, the computational complexity of the controller is also reduced and the communication resources are saved by introducing the command filters and event-triggered mechanism. It is proved that the tracking errors of the closed multi-agent systems are uniformly and ultimately bounded through Lyapunov theory. Finally, experimental simulations verify that the proposed control scheme is reasonable and effective. Furthermore, the proposed scheme provides a more flexible solution for containment control studies, especially for collision avoidance research of multi-agents.