Nighttime Traffic Light Location Based on Binocular Vision and Electronic Map

As online shopping is becoming mainstream, driven by the social impact of Coronavirus disease-2019 (COVID-19) as well as the development of Internet services, the demand for autonomous delivery mobile robots is rapidly increasing. This trend has brought the autonomous mobile robot market to a new turning point, with expectations that numerous mobile robots will be driving on roads with traffic. To achieve these expectations, autonomous mobile robots should precisely perceive the situation on roads with traffic. In this paper, we revisit and implement a real-time traffic light recognition system with a proposed lightweight state recognition network and ratio-preserving zero padding, which is a two-stage system consisting of a traffic light detection (TLD) module and a traffic light status recognition (TLSR) module. For the TLSR module, this work proposes a lightweight state recognition network with a small number of weight parameters, because the TLD module needs more weight parameters to find the exact location of traffic lights. Then, the proposed effective and lightweight network architecture is constructed by using skip connection, multifeature maps with different sizes, and kernels of appropriately tuned sizes. Therefore, the network has a negligible impact on the overall processing time and minimal weight parameters while maintaining high performance. We also propose to utilize a ratio-preserving zero padding method for data preprocessing for the TLSR module to enhance recognition accuracy. For the TLD module, extensive evaluations with varying input sizes and backbone network types are conducted, and then appropriate values for those factors are determined, which strikes a balance between detection performance and processing time. Finally, we demonstrate that our traffic light recognition system, utilizing the TLD module’s determined parameters, the proposed network architecture for the TLSR module, and the ratio-preserving zero padding method can reliably detect the location and state of traffic lights in real-world videos recorded in Gumi and Deagu, Korea, while maintaining at least 30 frames per second for real-time operation.

Traffic violations are common in the area zebra crossing at the location of traffic lights, violations are generally caused by the negligence of motorists who do not comply with existing regulations. As a result, there were many traffic accidents that could have been avoided. Starting from this problem, a simulation model tool is needed that is able to provide direct appeals to drivers when a violation occurs. Based on this, the purpose of this research is to create a simulation model to prevent accidents caused by motorists who are negligent of the rules by identifying vehicles that stop crossing the line boundaries. zebracross or drivers waiting for an out of place traffic light using imagery HaarCascade Classifiers which is programmed using Python and OpenCV to help process digital images. The results of this study that the Haar Cascade Classifier algorithm can characterize motorbikes well, with the best average accuracy value of 91.5% at a resolution of 720p. Based on the results obtained shows that the algorithm haar cascade classifier can detect violations that exist at traffic lights. Hence this indicates the rate of traffic violations can reduce by utilizing the algorithm haar cascade classifier.

Accurate localization and classification of traffic lights in driving scenes are crucial for enhancing road scene understanding in various intelligent vehicles applications. However, determining which traffic lights are relevant for the ego-vehicle remains an under-explored challenge. In this paper, we address both the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">local</i> task of identifying the state and relevance of each traffic light in an image and the strictly related <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">global</i> task of recommending the correct course of action for the ego-vehicle (should it stop?). We propose a novel architecture, which not only localizes each traffic light and identifies its relevance with respect to the ego-vehicle, but also generates a global recommendation. To address the scarcity of datasets with these types of annotations, we introduce the Verizon Connect Traffic Light Dataset (VZC-TLD), the first U.S. dataset that provides 3,000 images annotated with traffic light boxes, states, and relevance. Experimental results on both VZC-TLD and the DriveU Traffic Light Dataset (DTLD) show that our unified approach is indeed effective, and leads to significant improvements over approaches that do not exploit the synergies between the local and global tasks. Dataset is available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://leotac.github.io/vzc-tld</uri>

An automatic traffic light recognition system is proposed in this paper so that car drivers have sufficient information to make a correct decision. This in turn facilitates the construction of an ITS (Intelligent Transportation System). The proposed method can be applied to movable cameras, is adaptable to environmental variations and is proper for real‐time implementation. It consists of two phases: traffic light detection and classification. The two phases are based on color, region and border information. At the detection stage, the RGB color space is first converted into the HSI color space so as to find those regions with specific colors of traffic lights. The morphology technology is employed to remove holes and noise. Region labeling is then involved to detect candidate regions of traffic lights. At the classification stage, border detection is employed to obtain region borders which together with region colors are matching features. In this study, circle and arrow traffic lights can both be coped with. Moreover, our method is adaptable to environmental variations in numbers of lights, types of lights and time slots. Various experiments have been conducted to demonstrate the effectiveness and practicability of the proposed method.

Insufficient capacity of the road network is a problem in most cities. This problem can be partially solved by optimizing the operation of the traffic light. Since selecting traffic light settings in real time can lead to a deterioration in the traffic situation, it is advisable to pre-configure the model and work out various traffic light operating modes on it, and select the optimal ones. The purpose of the study was to simulate changes in the average number of cars in a queue before a traffic light and the average travel time of an intersection. To achieve this goal, the following tasks are formulated: creating a model of a road section with real flows at the entrances; determining the average number of cars in a queue under different traffic light operating modes; creating an experiment to find the average time of a left turn with and without the introduction of a specialized left-turn phase of the traffic light. Materials and methods. To solve the tasks set, the method of using the AnyLogic environment for analyzing a traffic section has been tested, and the possibility of using simulation modeling to identify the dependence of the intersection's capacity on the operating mode of the traffic light has been shown. Results. In the course of the research, a model of the existing intersection was created and the dependence of the number of cars in the queue in front of the forbidding traffic light signal on the duration of its full cycle was considered. To conduct the experiment, the existing 84-second traffic light cycle was changed to 30 seconds, one, two, three, and five minutes. To visualize the obtained dependencies, a graph is constructed that reflects the number of cars in a traffic jam under different traffic light operating modes. The created model is also used to determine the dependence of the travel time of an intersection at a left turn on the number of turning cars and the number of cars in the opposite direction. Traffic light options with and without an additional left signal are considered. The results are presented in a table and graphically. Conclusion. The study showed that with a small amount of oncoming traffic, the introduction of a specialized left-turn phase of the traffic light increases the travel time of the intersection, but with an increase in oncoming traffic, the situation changes to the opposite.

In this paper, we studied an autonomous vehicle driving system at an intersection equipped with traffic lights. We proposed deep learning based traffic light recognition algorithm, crossing/stop decision algorithm, path generation, and path following method. In addition, we implemented the proposed systems into a real vehicle environment. The main contribution of this paper is a crossing/stop decision algorithm to determine whether a test vehicle will pass through an intersection or stop in accordance with the recognized traffic light condition. This algorithm was designed based on vehicle speed and the distance between the position of the stop line and the vehicle. Ioniq PHEV, a vehicle equipped with sensors for autonomous driving, was used in order to evaluate the performance of the proposed system. The test vehicle drove through the intersection at the proving ground located in Ochang Campus, Chungbuk National University. The experimental results showed that the test vehicle successfully passed through the intersection in accordance with the traffic light status with a maximum speed of 30 kph on a straight course and a maximum speed of 10 kph on a 90° corner course.

This article presents research concerning the recognition of road traffic lights. Initially, vision algorithms were analysed regarding their suitability for implementation in vehicle control systems dedicated to individuals with specific communication needs. The paper presents the results of experimental studies on a vision system for recognising traffic lights, conducted using convolutional neural networks (CNNs). For the experiment, a custom database of traffic light images was prepared. This database was utilised to train a selected Xception CNN model and for processing by a classic algorithm based on colour analysis in the HSV colour space. The obtained classification accuracy results, reaching 98.75%, could serve as a 'green light' for implementing the developed technology to assist driving. The research findings may also find application in driver assistance systems, with particular attention given to the mobility of people with specific needs, such as those with visual impairments.

A return on investment analysis of applying health monitoring to LED lighting systems

In order to improve the city’s traffic efficiency, this paper presents a green wave optimal speed guidance system based on V2X (Vehicle-to-Everything) and the method to calculate the optimal velocity. All vehicles of the system should be equipped with V2X-Boxes which can broadcast the ID, time, location and speed of the vehicle while receiving the messages from other V2X-Boxes and the traffic light status from roadside unit (RSU) in real time. The V2X-Box calculates the optimal speed by acceleration guidance method and deceleration guidance method, which not only increases the number of vehicles passing through the green light, but also provides green wave optimal speed guidance for vehicles entering the arterial road from the bypass. This system guarantees all road-running vehicles through traffic lights without stopping.

The object of traffic route lighting in Britain is to enable vehicles to traverse built-up areas in safety, without the need for head lights. The technique adopted is to project light on to the road to show its direction and features, and to reveal clearly obstacles and other road users in silhouette. To achieve this an average road brightness of 0·2 foot-lamberts without excessive diversity is adequate. Glare from the lighting must be controlled to avoid disturbing a driver's visual adaptation but mainly to minimize his discomfort. A means of appraisal of the effectiveness of street lighting installations is described and applied to some installations. This should be of value to indicate which of the older installations are adequate and which need raising to a standard that affords an ample margin of safety without the use of head lights. Since it is left to the discretion of drivers to use head lights when they need them, it is suggested that the limitation in mounting height on other roads be relaxed sufficiently to enable lanterns to project safely over the road in order to effect considerable improvement in lighting, particularly on tree-lined streets.

A new traffic light detection and recognition algorithm for electronic travel aid was realized. Traffic lights were detected by rectangular and circular shapes extraction. Color distributions of the circular regions in HSI and RGB color space, combined with the position of the circular shape in the rectangle were used to recognize the traffic light. Gray level normalization, top hat transformation and image interception were used to reduce the disturbances of the environment. The average recognition rate is 97% and the speed is 0.086s per frame. Video analysis was used to improve the performance of the system.

Every year 1.3 million people worldwide are killed on roads, and between 20 and 50 million are injured. A good solution to this problem would be to develop machines, which take into account the environment. That is why today, safe auto driving is becoming a popular topic in many fields, from small projects to large car factories. However this topic also raises many questions and problems. There is a need to define the width of the edges of the road, recognize road signs, traffic lights, pedestrians, and other objects which contribute the driving safely. There are many methods for solving these tasks. In this thesis considered solution for detection edge of the road, objects on the road, traffic lights and traffic signs. Here are some of the methods which will be considered in the paper: Canny edge detection, threshold, Hough method and SURF -algorithm. During the calculations and experiments there were some mistakes and disadvantages which will be discussed below.

Traffic congestion has been a major problem for almost every city around the globe. Most of those cities get their traffic even worse in rush hour. In order to address the limitation of existing traffic light which is operated using fixed-time intervals, many researchers have been developing an intelligent traffic light that works in real time based on a machine learning algorithm. Unfortunately, it takes a lot of time to process the algorithm as the traffic is getting worse. In this paper, we propose a Q-Learning Accelerator Architecture for Intelligent Traffic Light Controller. It allows a faster learning time computation. Furthermore, the controller will allow traffic lights to work adaptively in real time. This architecture is compatible with a 4-arm crossroad using 4 possible actions which are traffic light signals operation. This architecture has been successfully implemented on PYNQ-ZI and achieve maximum clock frequency of 100 MHz. The achievable clock offers acceleration ratio approximately 1400 times against fully software Q-Learning on 2.3 GHz Intel Core i5 With 8 GB Memory. In addition, we also perform HW/SW co-verification by integrating HW design to the traffic simulator in order to validate system in realistic condition. Evaluation result shows that the proposed system offers better performance compare to conventional method, specifically can maintain the number of waiting vehicles. This result corresponds to reduction of congestion for around 75%.

In present-day the most common issues that concerns everyone is traffic congestion on road at crossings where traffic light poles are placed. The objective is to flow and maintain traffic system automatically but unfortunately jamming queue exists while everything is fine from technology point of view because of predefined static time duration. This predefined time slots fail to capture real time scenarios in terms of traffic load and that leads to congestion. To avoid such type of problem, this work contributes the concept of dynamic traffic light slots as a best possible solution. Dynamic traffic light is a concept of controlling the time duration of RED and GREEN light according to traffic flow or by sensing flow of traffic on road in movement. This can be achieved by using Global Positioning System which will convey the details of traffic flow and by considering that, time slots are provided to traffic lights i.e. whole controlling and managing will work like Real Time System.

In order to be able to make the image recognition and color recognition better fusion and application, this paper uses the YOLO v2 network that achieves excellent results in the target detection field, which can be through the training of traffic light samples to achieve the high speed and accuracy of identifying and positioning the traffic lights, in combination with the HSV color model, design the ratio of the red and green colors in the located area, identify the red and green colors, and then determine the status of the traffic lights. Through this study, we can understand the possibility of using YOLO v2 network in more fields, and at the same time combine color recognition to make YOLO v2 can get more extensive application.