Deep-Learning-Based Floor Path Model for Route Tracking of Autonomous Vehicles

Real-time route tracking is an important research topic for autonomous vehicles used in industrial facilities. Traditional methods such as copper line tracking on the ground, wireless guidance systems, and laser systems are still used in route tracking. In this study, a deep-learning-based floor path model for route tracking of autonomous vehicles is proposed. A deep-learning floor path model and algorithm have been developed for highly accurate route tracking, which avoids collisions of vehicles and follows the shortest route to reach the destination. The floor path model consists of markers. Routes in the floor path model are created by using these markers. The floor path model is transmitted to autonomous vehicles as a vector by a central server. The server dispatches the target marker address to the vehicle to move. The vehicle calculates all possible routes to this address and chooses the shortest one. Marker images on the selected route are processed using image processing and classified with a pre-trained deep-CNN model. If the classified image and the image on the selected route are the same, the vehicle proceeds toward its destination. While the vehicle moves on the route, it sends the last classified marker to the server. Other autonomous vehicles use this marker to determine the location of this vehicle. Other vehicles on the route wait to avoid a collision. As a result of the experimental studies we have carried out, the route tracking of the vehicles has been successfully achieved.

Intelligent Path Recognition against Image Noises for Vision Guidance of Automated Guided Vehicles in a Complex Workspace

Applying computer vision to mobile robot navigation has been studied more than twodecades. The most challenging problems for a vision-based AGV running in a complex workspaceinvolve the non-uniform illumination, sight-line occlusion or stripe damage, which inevitably resultin incomplete or deformed path images as well as many fake artifacts. Neither the fixed thresholdmethods nor the iterative optimal threshold methods can obtain a suitable threshold for the pathimages acquired on all conditions. It is still an open question to estimate the model parameters ofguide paths accurately by distinguishing the actual path pixels from the under- or oversegmentationerror points. Hence, an intelligent path recognition approach based on KPCA–BPNNand IPSO–BTGWP is proposed here, in order to resist the interferences from the complexworkspace. Firstly, curvilinear paths were recognized from their straight counterparts by means of apath classifier based on KPCA–BPNN. Secondly, an approximation method based on BTGWP wasdeveloped for replacing the curve with a series of piecewise lines (a polyline path). Thirdly, a robustpath estimation method based on IPSO was proposed to figure out the path parameters from a set ofpath pixels surrounded by noise points. Experimental results showed that our approach caneffectively improve the accuracy and reliability of a low-cost vision-guidance system for AGVs in acomplex workspace.

A Qualitative Analysis of a USB Camera for AGV Control

The increasing use of Automated Guided Vehicles (AGV) in the industry points to a search for better techniques and technologies to adapt to market requirements. Proper position control and movement give an AGV greater movement accuracy and greater lateral oscillations stability and vibration. It leads to smaller corridors and leaner plants, to more relaxed shipment devices, and to greater safety in the transport of fragile loads, for instance. AGV control techniques are not new, but new sensors’ applications are possible, such as USB cameras. In this sense, it is necessary to ensure the sensor is adequate to control system requirements. This work addresses AGVs driven by passive floor demarcations. It presents a qualitative analysis of a USB camera as sensors for AGV control, not yet a common industrial application. We performed the experiments with a small AGV prototype on an eight-shaped lane, varying both camera parameters and AGV parameters, such as linear speed. The AGV uses a USB camera with different image processing settings—different morphological filters structuring elements shapes and sizes, and three different image resolutions—to analyze the factors that affect line detection and control processing. This paper’s main contribution is a qualitative and quantitative analysis for the different sensor configurations. In addition, it discusses the influence sources on camera image as a position sensor. Furthermore, the experiments confirm sensor pertinence for the proposed control system.