Robust Navigation Research Articles

Bronchial branch tracing navigation in ultrathin bronchoscopy-guided radial endobronchial ultrasound for peripheral pulmonary nodule

BackgroundMost malignant peripheral pulmonary lesions (PPLs) are situated in the peripheral region of the lung. Although the ultrathin bronchoscope (UTB) can access these areas, a robust navigation system is essential for precise localisation of these small peripheral PPLs. Since many UTB procedures rely on automated virtual bronchoscopic navigation (VBN), this study aims to determine the accuracy and diagnostic yield of the manual bronchial branch tracing (BBT) navigation in UTB-guided radial endobronchial ultrasound (rEBUS) procedures.MethodsSingle-centre retrospective study of UTB-rEBUS patients with PPLs smaller than 3 cm over a two year period.ResultsOur cohort consisted of 47 patients with a mean age of 61.6 (SD 9.53) years and a mean target size of 1.91 (SD 0.53) cm. Among these lesions, 46.8% were located in the 6th airway generation, and 78.7% exhibited a direct bronchus sign. Navigation success using BBT was 91.5% based on positive rEBUS identification. The index diagnostic yield was 82.9%, increasing to 91.5% at 12 months of follow-up. Malignant lesions accounted for 65.1% of cases, while 34.9% were non-malignant. The presence of a direct bronchus sign was the sole factor associated with higher navigation success and diagnostic yield. Cryobiopsy outperformed forceps biopsy in non-concentric rEBUS lesions (90.9% vs. 50.0%, p < 0.05), but not in concentric orientated lesions. One pneumothorax occurred in our cohort.ConclusionsBBT as an exclusive navigation method for small PPLs in UTB-rEBUS procedures has proved to be safe and feasible. Combination of UTB with cryobiopsy remains efficient for eccentric and adjacently oriented rEBUS lesions.

Research on Multi-Source Data Fusion and Satellite Selection Algorithm Optimization in Tightly Coupled GNSS/INS Navigation Systems

With the increase in the number of Global Navigation Satellite System (GNSS) satellites and their operating frequencies, richer observation data are provided for the tightly coupled Global Navigation Satellite System/Inertial Navigation System (GNSS/INS). In this paper, we propose an efficient and robust combined navigation scheme to address the key issues of system accuracy, robustness, and computational efficiency. The tightly combined system fuses multi-source data such as the pseudo-range, the pseudo-range rate, and dual-antenna observations from the GNSS and the horizontal attitude angle from the vertical gyro (VG) in order to realize robust navigation in a sparse satellite observation environment. In addition, to cope with the high computational load faced by the system when the satellite observation conditions are good, we propose a weighted quasi-optimal satellite selection algorithm that reduces the computational burden of the navigation system by screening the observable satellites while ensuring the accuracy of the observation data. Finally, we comprehensively evaluate the proposed system through simulation experiments. The results show that, compared with the loosely coupled navigation system, our system has a significant improvement in state estimation accuracy and still provides reliable attitude estimation in regions with poor satellite observation conditions. In addition, in comparison experiments with the optimal satellite selection algorithm, our proposed satellite selection algorithm demonstrates greater advantages in terms of computational efficiency and engineering practicability.

Buoyancy Gliders Opening New Research Opportunities in the Southern Ocean

Abstract Polar systems are experiencing major changes that has significant implications for ocean circulation and global biogeochemistry. While these changes are accelerating, access to polar systems is decreasing as ships and logistical capabilities are declining. Autonomous underwater buoyancy gliders have proven to be robust technologies that are capable of filling sampling gaps. Gliders have also provided a more sustained presence in polar seas than ships are able. Along the West Antarctic Peninsula, one of the most rapidly warming regions on this planet, gliders have proven to be a useful tool being used by the international community to link land research stations without requiring major research vessel ship support. The gliders are capable of adaptive sampling of subsurface features not visible from satellites, sustained sampling to characterize seasonal dynamics, and they increasingly play a central role in the management of natural resources. Future challenges to expand their utility include: (A) developing robust navigation under ice, which would allow gliders to provide a sustained bridge between the research stations when ship support is declining, and (B) expanding online resources to provide the international community open access to quality data in near real time. These advances will accelerate the use of gliders to fill critical sampling gaps for these remote ocean environments.

Enhancing visual autonomous navigation in row-based crops with effective synthetic data generation

Abstract Introduction Service robotics is recently enhancing precision agriculture enabling many automated processes based on efficient autonomous navigation solutions. However, data generation and in-field validation campaigns hinder the progress of large-scale autonomous platforms. Simulated environments and deep visual perception are spreading as successful tools to speed up the development of robust navigation with low-cost RGB-D cameras. Materials and methods In this context, the contribution of this work resides in a complete framework to fully exploit synthetic data for a robust visual control of mobile robots. A wide realistic multi-crops dataset is accurately generated to train deep semantic segmentation networks and enabling robust performance in challenging real-world conditions. An automatic parametric approach enables an easy customization of virtual field geometry and features for a fast reliable evaluation of navigation algorithms. Results and conclusion The high quality of the generated synthetic dataset is demonstrated by an extensive experimentation with real crops images and benchmarking the resulting robot navigation both in virtual and real fields with relevant metrics.

Graph neural network based abnormal perception information reconstruction and robust autonomous navigation

PurposeThis paper aims to develop a robust navigation enhancement framework to handle one of the most urgent needs for real applications of autonomous vehicles nowadays, as these corner cases act as the most commonly occurred risks in potential self-driving accidents.Design/methodology/approachIn this paper, the main idea is to fully exploit the consistent features among spatio-temporal data and thus detect the anomalies and build residual channels to reconstruct the abnormal information. The authors first develop an anomaly detection algorithm, then followed by a corresponding disturbed information reconstruction network which has strong robustness to address both the nature disturbances and external attacks. Finally, the authors introduce a fully end-to-end resilient navigation performance enhancement framework to improve the driving performance of existing self-driving models under attacks and disturbances.FindingsComparison results on CARLA platform and real experiments demonstrate strong resilience of the authors’ approach which enhances the navigation performance under disturbances and attacks.Originality/valueReliable and resilient navigation performance under various nature disturbances and even external attacks is one of the most urgent needs for real applications of autonomous vehicles nowadays, as these corner cases act as the most commonly occurred risks in potential self-driving accidents. The information reconstruction approach provides a resilient navigation performance enhancement method for existing self-driving models.

Bi-directional adaptive enhanced A* algorithm for mobile robot navigation

PurposeThe present paper aims to address challenges associated with path planning and obstacle avoidance in mobile robotics. It introduces a pioneering solution called the Bi-directional Adaptive Enhanced A* (BAEA*) algorithm, which uses a new bidirectional search strategy. This approach facilitates simultaneous exploration from both the starting and target nodes and improves the efficiency and effectiveness of the algorithm in navigation environments. By using the heuristic knowledge A*, the algorithm avoids unproductive blind exploration, helps to obtain more efficient data for identifying optimal solutions. The simulation results demonstrate the superior performance of the BAEA* algorithm in achieving rapid convergence towards an optimal action strategy compared to existing methods.Design/methodology/approachThe paper adopts a careful design focusing on the development and evaluation of the BAEA* for mobile robot path planning, based on the reference [18]. The algorithm has remarkable adaptability to dynamically changing environments and ensures robust navigation in the context of environmental changes. Its scale further enhances its applicability in large and complex environments, which means it has flexibility for various practical applications. The rigorous evaluation of our proposed BAEA* algorithm with the Bidirectional adaptive A* (BAA*) algorithm [18] in five different environments demonstrates the superiority of the BAEA* algorithm. The BAEA* algorithm consistently outperforms BAA*, demonstrating its ability to plan shorter and more stable paths and achieve higher success rates in all environments.FindingsThe paper adopts a careful design focusing on the development and evaluation of the BAEA* for mobile robot path planning, based on the reference [18]. The algorithm has remarkable adaptability to dynamically changing environments and ensures robust navigation in the context of environmental changes. Its scale further enhances its applicability in large and complex environments, which means it has flexibility for various practical applications. The rigorous evaluation of our proposed BAEA* algorithm with the Bi-directional adaptive A* (BAA*) algorithm [18] in five different environments demonstrates the superiority of the BAEA* algorithm.Research limitations/implicationsThe rigorous evaluation of our proposed BAEA* algorithm with the BAA* algorithm [18] in five different environments demonstrates the superiority of the BAEA* algorithm. The BAEA* algorithm consistently outperforms BAA*, demonstrating its ability to plan shorter and more stable paths and achieve higher success rates in all environments.Originality/valueThe originality of this paper lies in the introduction of the bidirectional adaptive enhancing A* algorithm (BAEA*) as a novel solution for path planning for mobile robots. This algorithm is characterized by its unique characteristics that distinguish it from others in this field. First, BAEA* uses a unique bidirectional search strategy, allowing to explore the same path from both the initial node and the target node. This approach significantly improves efficiency by quickly converging to the best paths and using A* heuristic knowledge. In particular, the algorithm shows remarkable capabilities to quickly recognize shorter and more stable paths while ensuring higher success rates, which is an important feature for time-sensitive applications. In addition, BAEA* shows adaptability and robustness in dynamically changing environments, not only avoiding obstacles but also respecting various constraints, ensuring safe path selection. Its scale further increases its versatility by seamlessly applying it to extensive and complex environments, making it a versatile solution for a wide range of practical applications. The rigorous assessment against established algorithms such as BAA* consistently shows the superior performance of BAEA* in planning shorter paths, achieving higher success rates in different environments and cementing its importance in complex and challenging environments. This originality marks BAEA* as a pioneering contribution, increasing the efficiency, adaptability and applicability of mobile robot path planning methods.

Reinforcement learning based robot navigation using illegal actions for autonomous docking of surface vehicles in unknown environments

Autonomous Surface Vehicles (ASVs) are bound to play a fundamental role in the maintenance of offshore wind farms. Robust navigation for inspection vehicles should take into account the operation of docking within a harbouring structure, which is a critical and still unexplored maneuver. This work proposes an end-to-end docking approach for ASVs, based on Reinforcement Learning (RL), which teaches an agent to tackle collision-free navigation towards a target pose that allows the berthing of the vessel. The developed research presents a methodology that introduces the concept of illegal actions to facilitate the vessel’s exploration during the learning process. This method improves the adopted Actor-Critic (AC) framework by accelerating the agent’s optimization by approximately 38.02%. A set of comprehensive experiments demonstrate the accuracy and robustness of the presented method in scenarios with simulated environmental constraints (Beaufort Scale and Douglas Sea Scale), and a diversity of docking structures. Validation with two different real ASVs in both controlled and real environments demonstrates the ability of this method to enable safe docking maneuvers without prior knowledge of the scenario.

Development of Blind Campus Navigation System with Obstacle Detection Device

This study focuses on integrating Text-To-Speech software, Global Positioning System (GPS) and other technologies attached to existing white cane to create a robust navigation system that provides real-time feedback and assistance to Students with Visual Impairment (SVI) using Nigerian accent. It uses the design science research methodology for the development and validation of the GPS based mobility into object detection white cane for orientation and mobility of SVI. A speech-corpus database was created to serve as a dictionary for the Text-To-Speech and synthesized through machine learning and artificial intelligence to enable the object detection white cane to detect objects and identify common places at 30 meters in Federal College of Education (Special), Oyo campus, Oyo state, Nigeria. The developed object detection white cane was evaluated with 20 SVI selected for the study using the purposive sampling technique and data were collected through interviews and questionnaires. Two research questions were raised for the study. Data collected were analyzed both quantitatively and qualitatively, using Statistical Package for the Social Sciences (SPSS) and Atlas.ti. The results revealed that the mean response of the participants to all the items on the integration of Text-To-Speech software into object detection white cane is “1” an indication that Text-To-Speech software enhances the independent navigation of students with visual impairment. The study recommended that the components used were imported and expensive, hence the need for locally source components that can be used in producing the devices in large quantities and at reduced cost.

A bio-inspired model for robust navigation assistive devices

This paper proposes a new implementation and evaluation in a real-world environment of a bio-inspired predictive navigation model for mobility control, suitable especially for assistance of visually impaired people and autonomous mobile systems. This bio-inspired model relies on the interactions between formal models of three types of neurons identified in the mammals’ brain implied in navigation tasks, namely place cells, grid cells, and head direction cells, to construct a topological model of the environment under the form of a decentralized navigation graph. Previously tested in virtual environments, this model demonstrated a high tolerance to motion drift, making possible to map large environments without the need to correct it to handle such drifts, and robustness to environment changes. The presented implementation is based on a stereoscopic camera, and is evaluated on its possibilities to map and guide a person or an autonomous mobile robot in an unknown real environment. The evaluation results confirm the effectiveness of the proposed bio-inspired navigation model to build a path map, localize and guide a person through this path. The model predictions remain robust to environment changes, and allow to estimate traveled distances with an error rate below 3% over test paths, up to 100m. The tests performed on a robotic platform also demonstrated the pertinence of navigation data produced by this navigation model to guide an autonomous system. These results open the way toward efficient wearable assistive devices for visually impaired people independent navigation.

Angle Robustness Unmanned Aerial Vehicle Navigation in GNSS-Denied Scenarios

Due to the inability to receive signals from the Global Navigation Satellite System (GNSS) in extreme conditions, achieving accurate and robust navigation for Unmanned Aerial Vehicles (UAVs) is a challenging task. Recently emerged, vision-based navigation has been a promising and feasible alternative to GNSS-based navigation. However, existing vision-based techniques are inadequate in addressing flight deviation caused by environmental disturbances and inaccurate position predictions in practical settings. In this paper, we present a novel angle robustness navigation paradigm to deal with flight deviation in point-to-point navigation tasks. Additionally, we propose a model that includes the Adaptive Feature Enhance Module, Cross-knowledge Attention-guided Module and Robust Task-oriented Head Module to accurately predict direction angles for high-precision navigation. To evaluate the vision-based navigation methods, we collect a new dataset termed as UAV_AR368. Furthermore, we design the Simulation Flight Testing Instrument (SFTI) using Google Earth to simulate different flight environments, thereby reducing the expenses associated with real flight testing. Experiment results demonstrate that the proposed model outperforms the state-of-the-art by achieving improvements of 26.0% and 45.6% in the success rate of arrival under ideal and disturbed circumstances, respectively.

Optimizing AUV Navigation Using Factor Graphs with Side-Scan Sonar Integration

For seabed mapping, the prevalence of autonomous underwater vehicles (AUVs) employing side-scan sonar (SSS) necessitates robust navigation solutions. However, the positioning errors of traditional strapdown inertial navigation system (SINS) and Doppler velocity log (DVL) systems accumulated significantly, further exacerbated by DVL’s susceptibility to failure in complex underwater conditions. This research proposes an integrated navigation approach that utilizes factor graph optimization (FGO) along with an improved pre-integration technique integrating SSS-derived position measurements. Firstly, the reliability of SSS image registration in the presence of strong noise and feature-poor environments is improved by replacing the feature-based methods with a Fourier-based method. Moreover, the high-precision inertial measurement unit (IMU) pre-integration method could correct the heading errors of SINS significantly by considering the Earth’s rotation. Finally, the AUV’s marine experimental results demonstrated that the proposed integration method not only offers improved SSS image registration and corrects initial heading discrepancies but also delivers greater system stability, particularly in instances of DVL data loss.

Robust Strategy Generation for Automatic Navigation of Unmanned Surface Vehicles through Improved DDPG Algorithm

Automatic navigation with collision‐free navigation has become a critical challenge for unmanned surface vehicles (USVs) to expand their application scenarios. Conventional methods for achieving automatic navigation of USVs typically rely on finely modeling the environment, thus exhibiting poor generalization capabilities. Methods based on deep reinforcement learning possess powerful learning abilities and have achieved promising results in USV‐automatic navigation‐tasks. However, the increase in the complexity of network model structures has led to instability during the training process. Therefore, generating more robust navigation strategies, namely ensuring robust reward‐score trends during training and smoother action trajectories of the USV, is crucial for automatic navigation and constitutes the main research question of this study. In this paper, an improved deep deterministic policy gradient (DDPG) algorithm has been proposed for stable automatic navigation of USVs in complex environments. In this algorithm, first, we construct a stable training framework that incorporates the stable feature‐sharing module with constrained gradient backpropagation, which bolsters the USV’s scene memorization capacity, reduces model training fluctuations during navigation policy learning, and improves the training stability of the navigation model. Second, we ensure the decision adaptability of the USV by constraining the extent of action change between adjacent time‐steps by using a reward‐function, which improves the USV‐action smoothly. Finally, we design typical USV‐automatic‐navigation‐scenarios to validate the performance of the Algorithm. Experimental results validate our algorithm’s capability to achieve collision‐free navigation, outperforming the traditional DDPG algorithm in terms of convergence speed, effective sailing distance, and rudder angle maneuver consumption, among other performance metrics.

Learning Heterogeneous Relation Graph and Value Regularization Policy for Visual Navigation.

The goal of visual navigation is steering an agent to find a given target object with current observation. It is crucial to learn an informative visual representation and robust navigation policy in this task. Aiming to promote these two parts, we propose three complementary techniques, heterogeneous relation graph (HRG), a value regularized navigation policy (VRP), and gradient-based meta learning (ML). HRG integrates object relationships, including object semantic closeness and spatial directions, e.g., a knife is usually co-occurrence with bowl semantically or located at the left of the fork spatially. It improves visual representation learning. Both VRP and gradient-based ML improve robust navigation policy, regulating this process of the agent to escape from the deadlock states such as being stuck or looping. Specifically, gradient-based ML is a type of supervision method used in policy network training, which eliminates the gap between the seen and unseen environment distributions. In this process, VRP maximizes the transformation of the mutual information between visual observation and navigation policy, thus improving more informed navigation decisions. Our framework shows superior performance over the current state-of-the-art (SOTA) in terms of success rate and success weighted by length (SPL). Our HRG outperforms the Visual Genome knowledge graph on cross-scene generalization with ≈ 56% and ≈ 39% improvement on Hits@ 5* (proportion of correct entities ranked in top 5) and MRR * (mean reciprocal rank), respectively. Our code and HRG datasets will be made publicly available in the scientific community.

Mechanization of a Robust Navigation Scheme for Land Applications Employing GNSS/INS Augmented with Zero Velocity Update

Mechanization of a Robust Navigation Scheme for Land Applications Employing GNSS/INS Augmented with Zero Velocity Update

A NOVEL GPS FAULT DETECTION AND EXCLUSION ALGORITHM AIDED BY IMU AND VO DATA FOR VEHICLE INTEGRATED NAVIGATION IN URBAN ENVIRONMENTS

Abstract. High accuracy positioning is crucial for various applications that require accurate and reliable positioning data, such as autonomous vehicles (AVs), agriculture, and intelligent transportation. Global Positioning System (GPS) is widely used in integration with Inertial Measurement Unit (IMU) and Visual Odometry (VO) to implement an accurate and robust navigation system. However, the performance of the integrated system can be severely degraded in urban environments due to non-line-of-sight (NLOS) reception and multipath interference. To overcome these challenges, a novel Fault Detection and Exclusion (FDE) algorithm aided with IMU and VO measurements is proposed to identify and isolate the contaminated satellite signals. The algorithm utilizes instantaneous measurements from IMU and VO to predict the current vehicle position, enabling the estimation of pseudorange errors in GPS measurements. A FDE algorithm based on Hierarchical clustering is then developed to identify GPS signals with significant errors based on the predicted pseudorange error. An experimental field test was conducted using a land vehicle to evaluate the effectiveness of the proposed algorithm. The results show that the GPS/IMU/VO integrated navigation system with the proposed FDE algorithm has significantly improved the positioning accuracy and reliability compared to the traditional system. The proposed algorithm achieves a positioning accuracy with a 3D Root Mean Square Error (RMSE) of 11.18m in urban environment, making an improvement of 68.2% over the traditional GPS/IMU/VO integrated navigation system.

Visual Navigation Subject to Embodied Mismatch

In the embodied visual navigation task, the agent navigates to a target location based on the visual observation it collects during the interaction with the environment. And various approaches have been proposed to learn robust navigation strategies for this task. However, existing approaches assume that the action spaces in the training and testing phases are the same, which is usually not the case in reality. And thus it is difficult to directly apply these approaches on practical scenarios. In this paper, we consider the situation where the action spaces in the training and testing phases are different, and a novel task of visual navigation subject to embodied mismatch is proposed. To solve the proposed task, we establish a two-stage robust adversary learning framework which can learn a robust policy to adapt the learned model to a new action space. In the first stage, an adversary training mechanism is used to learn a robust feature representation of the state. In the second stage, an adaptation training is used to transfer the learned strategy to a new action space with fewer training samples. Experiments of three types of embodied visual navigation tasks are conducted in 3D indoor scenes demonstrating the effectiveness of the proposed approach.

MVSSC: Meta-reinforcement learning based visual indoor navigation using multi-view semantic spatial context

In Visual Indoor Navigation (VIN), Deep Reinforcement Learning (DRL) is commonly used by agents to achieve end-to-end mapping from vision to action when navigating toward a target based on observation. However, current DRL-based work suffers from two challenges: partial observability resulting from using solely a single first-person view and poor generalization in the case of unknown scenes and unknown objects. In light of these issues, this paper introduces the integration of multi-view as an expansion of observability and meta-learning as a primary generalization technique into the DRL framework and presents the meta-reinforcement learning method that leverages Multi-View Semantic Spatial Context (MVSSC). Specifically, aiming to explore the informative multi-view context for better efficiency in searching for and navigating to the target, we model the objects’ relationship from two aspects, multi-view semantic context (MVSEC) and multi-view spatial context (MVSPC). MVSEC enables agents to encode prior semantic relationships adaptively via a multi-view modulated graph. Meanwhile, MVSPC enhances the spatial representation of target-related objects’ correlation through similarity grids of multi-view. After adaptively fusing the multi-view and context information under the meta-reinforcement learning framework, our method can encourage efficient target search and robust navigation with stronger generalization performance to unknown scenes and unknown objects. Extensive experimental results on the AI2-THOR simulator demonstrate that our method outperforms the current state-of-the-art approaches.

Efficient air pollution mapping in extensive regions with fully autonomous unmanned aerial vehicles: A numerical perspective

It is noteworthy that comprehensive exploration of atmospheric measurements in the horizontal plane using aerial platforms, necessitating high autonomy, has not been extensively covered in the existing literature. This research presents a systematic numerical approach to effective air pollution mapping achieved through the integration of horizontal and vertical air pollution measurements conducted using a fully autonomous unmanned aerial vehicle (UAV) platform. The developed robust navigation model enables the UAV to efficiently scan the extensive measurement area, which is subdivided into smaller sub-areas using the polygonal decomposition technique, resulting in a comprehensive map of the entire region. Furthermore, technical analysis determines the optimal flight speed, leading to air pollution measurements in up to 30 % more areas and ensuring more consistent results. The simulation results illustrate the effective mapping of the entire area by aggregating air pollution measurements from sub-areas, with seamless transitions emphasizing the accuracy and consistency of the employed air pollution mapping technique. This systematic method offers numerous advantages, including rapid air pollution source identification and swift response capabilities. Moreover, this approach holds potential for various applications, such as forest fire monitoring and natural resources assessment, by equipping UAVs with additional equipment like cameras alongside atmospheric sensors.

ArthroNavi framework: stereo endoscope-guided instrument localization for arthroscopic minimally invasive surgeries.

As an example of a minimally invasive arthroscopic surgical procedure, arthroscopic osteochondral autograft transplantation (OAT) is a common option for repairing focal cartilage defects in the knee joints. Arthroscopic OAT offers considerable benefits to patients, such as less post-operative pain and shorter hospital stays. However, performing OAT arthroscopically is an extremely demanding task because the osteochondral graft harvester must remain perpendicular to the cartilage surface to avoid differences in angulation. We present a practical ArthroNavi framework for instrument pose localization by combining a self-developed stereo endoscopy with electromagnetic computation, which equips surgeons with surgical navigation assistance that eases the operational constraints of arthroscopic OAT surgery. A prototype of a stereo endoscope specifically fit for a texture-less scene is introduced extensively. Then, the proposed framework employs the semi-global matching algorithm integrating the matching cubes method for real-time processing of the 3D point cloud. To address issues regarding initialization and occlusion, a displaying method based on patient tracking coordinates is proposed for intra-operative robust navigation. A geometrical constraint method that utilizes the 3D point cloud is used to compute a pose for the instrument. Finally, a hemisphere tabulation method is presented for pose accuracy evaluation. Experimental results show that our endoscope achieves 3D shape measurement with an accuracy of . The mean error of pose localization is 15.4deg (range of 10.3deg to 21.3deg; standard deviation of 3.08deg) in our ArthroNavi method, which is within the same order of magnitude as that achieved by experienced surgeons using a freehand technique. The effectiveness of the proposed ArthroNavi has been validated on a phantom femur. The potential contribution of this framework may provide a new computer-aided option for arthroscopic OAT surgery.

Integrated robust navigation and guidance for the kinetic impact of near-earth asteroids based on deep reinforcement learning

The defense against near-Earth asteroids (NEAs) using kinetic impact is faced with various challenges, including limited maneuverability of the impactor, inaccurate dynamic models, poor observability of relative navigation, and control execution errors. To address these challenges, an integrated robust navigation and guidance method for the kinetic impact of NEAs based on deep reinforcement learning (DRL) is proposed in this paper, which can directly map angle measurements by a monocular camera to guidance maneuvers. Firstly, a discrete partial observable Markov decision process (POMDP) is modeled for the integrated navigation and guidance problem of NEA interception. To address the situation where impactors are usually only equipped with aiming cameras and only the line of sight (LOS) angle is measured, past and current LOS measurements are concatenated into a one-dimensional state observation vector to directly introduce the memory of historical state information. A shaping reward function design based on potential energy has also been proposed to solve the problem of sparse reward as the main goal. Subsequently, proximal policy optimization (PPO) is used to solve the established POMDP model, obtaining an integrated navigation and guidance policy that directly maps from the original output of the navigation sensor to the guidance command. The potentially hazard asteroid (PHA) Bennu is used as the target and a typical kinetic impact defense scenario is designed to simulate and verify the model and method proposed in this paper, considering the influence of multiple factors. Numerical simulation results show that the proposed method achieves an interception accuracy of 203.58 m (mean value). The proposed method abandons the traditional separate design of navigation and guidance algorithms, and its robustness has been tested and verified in a wide range of uncertain environments.