Deep Learning-Based Complete Coverage Path Planning With Re-Joint and Obstacle Fusion Paradigm

Lei, Tingjun; Luo, Chaomin; Jan, Gene Eu; Bi, Zhuming

doi:10.3389/frobt.2022.843816

Cited by 19 publications

(12 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The major function of the detection robot is to precisely and efficiently search the whole broiler barn and indicate the position of dead broilers. Thus, coverage path planning (CPP) algorithms need to be introduced to assist robots in search and exploration tasks, such as bio-inspired neural networks [22] , Boustrophedon grid decomposition [23] , deep learning [24,25] , etc.…”

Section: Related Workmentioning

confidence: 99%

“…Similarly, Nasirian et al utilized traditional graph theory to segment the workspace and proposed a deep reinforcement learning approach to solve the CCPP problem in a complex workspace [24] . Lei et al proposed a deep learning method to detect the workspace and generate turning waypoints for the robot to complete the coverage of the entire workspace [25] . The trajectory generated by the above-presented algorithms completely covers the workspace in light of the size of the robot, which is more suitable for a largesized workspace with a broad sensing range.…”

Section: Related Workmentioning

confidence: 99%

“…As a daily task, producers need to collect and remove broiler mortality timely, which is an arduous, time-consuming, and unpleasant task. A farmer spends hours daily identifying, gathering, and transporting dead birds in a typical broiler barn (e.g., 25,000 birds/barn). Manual mortality collection is relatively simple when broilers are young because they are lightweight and small.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An informative planning-based multi-layer robot navigation system as applied in a poultry barn

Lei¹,

Li²,

Luo³

et al. 2022

Intell Robot

View full text Add to dashboard Cite

Many real-world robot applications, as found in precision agriculture, poultry farms, disaster response, and environment monitoring, require search, locate, and removal (SLR) operations by autonomous mobile robots. In such application settings, the robots initially search and explore the entire workspace to find the targets, so that the subsequent robots conveniently move directly to the targets to fulfill the task. A multi-layer robot navigation system is necessary for SLR operations. The scenario of interest is the removal of broiler mortality by autonomous robots in poultry barns in this paper. Daily manual collection of broiler mortality is time- and labor-consuming, and an autonomous robotic system can solve this issue effectively. In this paper, a multi-layer navigation system is developed to detect and remove broiler mortality with two robots. One robot is assigned to search a large-scale workspace in a coverage mode and find and locate objects, whereas the second robot directly moves to the located targets to remove the objects. Directed coverage path planning (DCPP) fused with an informative planning protocol (IPP) is proposed to efficiently search the entire workspace. IPP is proposed for coverage directions in DCPP devoted to rapidly achieving spatial coverage with the least estimation uncertainty in the decomposed grids. The detection robot consists of a developed informative-based directed coverage path planner and a You Only Look Once (YOLO) V4-based dead bird detector. It refines and optimizes the coverage path based on historical data on broiler mortality distribution in a broiler barn. The removal robot collects dead broilers driven by a new hub-based multi-target path routing (HMTR) scheme, which is applicable to row-based environments. The proposed methods show great potential to navigate in broiler barns efficiently and safely, thus being a useful component for robotics. The effectiveness and robustness of the proposed methods are validated through simulation and comparison studies.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An informative planning-based multi-layer robot navigation system as applied in a poultry barn

Lei¹,

Li²,

Luo³

et al. 2022

Intell Robot

View full text Add to dashboard Cite

show abstract

“…Robotics system has been applied to numerous fields, such as transportation [1] , healthcare service [2,3] , agriculture [4] , manufacturing [5] , etc., in recent years. Robot navigation is one of the fundamental components in robotic systems, which includes multi-waypoint navigation system [6][7][8] .…”

Section: Introductionmentioning

confidence: 99%

A node selection algorithm to graph-based multi-waypoint optimization navigation and mapping

Sellers¹,

Lei²,

Luo³

et al. 2022

Intell Robot

View full text Add to dashboard Cite

Autonomous robot multi-waypoint navigation and mapping have been demanded in many real-world applications found in search and rescue (SAR), environmental exploration, and disaster response. Many solutions to this issue have been discovered via graph-based methods in need of switching the robotos trajectory between the nodes and edges within the graph to create a trajectory for waypoint-to-waypoint navigation. However, studies of how waypoints are locally bridged to nodes or edges on the graphs have not been adequately undertaken. In this paper, an adjacent node selection (ANS) algorithm is developed to implement such a protocol to build up regional path from waypoints to nearest nodes or edges on the graph. We propose this node selection algorithm along with the generalized Voronoi diagram (GVD) and Improved Particle Swarm Optimization (IPSO) algorithm as well as a local navigator to solve the safety-aware concurrent graph-based multi-waypoint navigation and mapping problem. Firstly, GVD is used to form a Voronoi diagram in an obstacle populated environment to construct safety-aware routes. Secondly, the sequence of multiple waypoints is created by the IPSO algorithm to minimize the total travelling cost. Thirdly, while the robot attempts to visit multiple waypoints, it traverses along the edges of the GVD to plan a collision-free trajectory. The regional path from waypoints to the nearest nodes or edges needs to be created to join the trajectory by the proposed ANS algorithm. Finally, a sensor-based histogram local reactive navigator is adopted for moving obstacle avoidance while local maps are constructed as the robot moves. An improved -spline curve-based smooth scheme is adopted that further refines the trajectory and enables the robot to be navigated smoothly. Simulation and comparison studies validate the effectiveness and robustness of the proposed model.

show abstract

“…Recently, the DeepWay method [24] has been proposed to efficiently combine deep learning and clustering for the generation of start and end row waypoints given an occupancy grid of the vineyard. Moreover, novel contributions adopted the same paradigm and training procedure to extend the coverage to arbitrary unstructured environments [17]. Despite being an important baseline for rowbased path generation, DeepWay leaves substantial space for improvement.…”

Section: Introductionmentioning

confidence: 99%

Waypoint Generation in Row-based Crops with Deep Learning and Contrastive Clustering

Salvetti¹,

Angarano²,

Martini³

et al. 2022

Preprint

View full text Add to dashboard Cite

The development of precision agriculture has gradually introduced automation in the agricultural process to support and rationalize all the activities related to field management. In particular, service robotics plays a predominant role in this evolution by deploying autonomous agents able to navigate in fields while executing different tasks without the need for human intervention, such as monitoring, spraying and harvesting. In this context, global path planning is the first necessary step for every robotic mission and ensures that the navigation is performed efficiently and with complete field coverage. In this paper, we propose a learning-based approach to tackle waypoint generation for planning a navigation path for row-based crops, starting from a top-view map of the region-of-interest. We present a novel methodology for waypoint clustering based on a contrastive loss, able to project the points to a separable latent space. The proposed deep neural network can simultaneously predict the waypoint position and cluster assignment with two specialized heads in a single forward pass. The extensive experimentation on simulated and real-world images demonstrates that the proposed approach effectively solves the waypoint generation problem for both straight and curved row-based crops, overcoming the limitations of previous state-of-the-art methodologies.

show abstract

Deep Learning-Based Complete Coverage Path Planning With Re-Joint and Obstacle Fusion Paradigm

Cited by 19 publications

References 45 publications

An informative planning-based multi-layer robot navigation system as applied in a poultry barn

An informative planning-based multi-layer robot navigation system as applied in a poultry barn

A node selection algorithm to graph-based multi-waypoint optimization navigation and mapping

Waypoint Generation in Row-based Crops with Deep Learning and Contrastive Clustering

Contact Info

Product

Resources

About