To Go or Not to Go: Shedding Light on Traffic Light Signal Manipulation and Defense Strategies

Traffic light and sign detectors on autonomous cars are integral for road scene perception. The literature is abundant with deep learning networks that detect either lights or signs, not both, which makes them unsuitable for real-life deployment due to the limited graphics processing unit (GPU) memory and power available on embedded systems. The root cause of this issue is that no public dataset contains both traffic light and sign labels, which leads to difficulties in developing a joint detection framework. We present a deep hierarchical architecture in conjunction with a mini-batch proposal selection mechanism that allows a network to detect both traffic lights and signs from training on separate traffic light and sign datasets. Our method solves the overlapping issue where instances from one dataset are not labelled in the other dataset. We are the first to present a network that performs joint detection on traffic lights and signs. We measure our network on the Tsinghua-Tencent 100K benchmark for traffic sign detection and the Bosch Small Traffic Lights benchmark for traffic light detection and show it outperforms the existing Bosch Small Traffic light state-of-the-art method. We focus on autonomous car deployment and show our network is more suitable than others because of its low memory footprint and real-time image processing time. Qualitative results can be viewed at https://youtu.be/ YmogPzBXOw.

Traffic signal lights system is a signalling device located an intersection or pedestrian crossing to control the movement of traffic. The timing of traffic signal lights has attracted many researchers to study the problems involving traffic light management and looking for an inexpensive and effective solution that requires inexpensive changes in the infrastructures. A simple traffic lights system uses a pre-timed control setting based on the latest traffic data, and the setting could be manually changed. It is a common type of signal control and sometimes the setting was not correctly configured with the traffic data, thus leading to congestion at an intersection. Many mathematical strategies were applied to get an optimal setting. This study aims to model the traffic flow at Persiaran Kayangan and Persiaran Permai Intersection, Section 7, Shah Alam, as the case study, by using AnyLogic simulation software. The model was used to determine the best timings of traffic green lights that minimise the average time at the intersection and reduce traffic congestion. The findings showed that the best timings of traffic green lights for four directions at the intersection are 120 seconds, 75 seconds, 130 seconds and 100 seconds, respectively. These timings of green lights produced the lowest average time at the intersection (55.65 seconds).

Traffic signal lights recognition system is an essential part of Advanced Driver Assistance Systems (ADAS). Methods for traffic lights recognition based on single feature and fixed threshold filtering are usually ineffective in complex background and variable lighting environment. To solve this problem, an approach based on features combination of color and shape of traffic lights is proposed, and the method of machine learning is used for recognition of traffic lights. On the basis of extracting the characteristic parameters of the candidate region, the SVM classifier is constructed to classify the traffic signal lights. Experimental results show that this method can realize the accurate location and recognition of traffic lights in complex scenes.

From the functional point of view, the intersection is the most complex element of the road network. It is here that the traffic flows in different directions cross, and various maneuvers take place. This indicates that the intersection is a place with an increased concentration of conflict situations and an increased risk of traffic accidents. At most high-flow intersections, traffic is controlled by traffic lights, and their inefficient operation can lead to unnecessarily long wait times and overall increase in traffic delays. The research analyzed a regulated intersection at the crossing of Stepana Bandery and Viacheslava Chornovola main streets in the city of Rivne. This intersection is characterized by constant significant traffic jams. To simulate traffic conditions at this intersection, the research used PTV Vissim software. Observations of traffic flows, their distribution by directions, and traffic light regulation parameters were used as initial data for modeling the intersection traffic scheme. The main problem that must be solved when considering isolated traffic light intersections is the calculation and optimization of the signal plan, which involves: determining the number of phases; determining cycle duration; distribution, i.e. determining parts of available green time for each phase; modeling of traffic situations that may arise due to the passage of priority vehicles, congestion of vehicles in intersection areas during peak periods or other situations. At the same time, it is necessary to achieve the best possible characteristics of the intersection functioning. According to the research results, the road conditions were modeled and the signal plan was optimized with the use of PTV Vissim software. Simulation modeling of the intersection included drawing a road network, installing traffic lights (signal controllers) with a description of their work (choosing the type of light signaling devices, creating signal groups and traffic light signals, parameters for coordinating signals), forming pedestrian zones and a node, performing calculations with subsequent analysis of the received data. In order to improve the efficiency of the intersection, two options for the operation of the traffic light controllers are offered: 1. Lengthening of the “green” phase, for the convenience of turning to the left (by reducing the phase of oncoming traffic in one direction by 5 seconds). The total duration of the cycle of 70 seconds will not change; 2. Implementation of the third phase – fully pedestrian in all directions, lasting 20 seconds. The total duration of the cycle will increase to 90 seconds. A more progressive measure can be the introduction of adaptive systems, which are based on new traffic monitoring technologies and allow obtaining accurate data on traffic flows in real time and performing adaptive control of traffic lights, that is, adapting the signal plan in real time to changes in traffic flows.

This work aims to implement traffic light and sign detection using Image processing technique for an autonomous and vehicle. Traffic Sign Recognition system is used to regulate traffic signs, warn a driver and command certain actions. Fast robust and real-time automatic traffic sign detection and recognition can support the driver and significantly increase driving safety. Automatic recognition of traffic signs is also important for an automated intelligent driving vehicle or for a driver assistance system. This is a visual based project i.e., the input to the system is video data which is continuously captured from the webcam is interfaced to the Raspberry Pi. Images are pre-processed with several image processing techniques such as; Hue, Saturation and Value (HSV) color space model technique is employed for traffic light detection, for sign detection again HSV color space model and Contour Algorithm has been used. The signs are detected based on Region of Interest (ROI). The ROI is detected based on the features like geometric shape and color of the object in the image containing the traffic signs. The experimental results show highly accurate classifications of traffic sign patterns with complex background images as well as the results accomplish in reducing the computational cost of this proposed method.

Autonomous vehicles are a promising solution to traffic congestion, air pollution, accidents, wasted time, and resources. However, remote driver intervention may be necessary in extreme situations to ensure safe roadside parking or complete remote takeover. In these cases, high-quality real-time video streaming is crucial for remote driving. In a preliminary study, we presented a region of interest (ROI) High-Efficiency Video Coding (HEVC) method where the image was segmented into two categories: ROI and background. This involved allocating more bandwidth to the ROI, which yielded an improvement in the visibility of classes essential for driving while transmitting the background at a lower quality. However, migrating the bandwidth to the large ROI portion of the image did not substantially improve the quality of traffic signs and lights. This study proposes a method that categorizes ROIs into three tiers: background, weak ROI, and strong ROI. To evaluate this approach, we utilized a photo-realistic driving scenario database created with the Cognata self-driving car simulation platform. We used semantic segmentation to categorize the compression quality of a Coding Tree Unit (CTU) according to its pixel classes. A background CTU contains only sky, trees, vegetation, or building classes. Essentials for remote driving include classes such as pedestrians, road marks, and cars. Difficult-to-recognize classes, such as traffic signs (especially textual ones) and traffic lights, are categorized as a strong ROI. We applied thresholds to determine whether the number of pixels in a CTU of a particular category was sufficient to classify it as a strong or weak ROI and then allocated bandwidth accordingly. Our results demonstrate that this multi-category ROI compression method significantly enhances the perceptual quality of traffic signs (especially textual ones) and traffic lights by up to 5.5 dB compared to a simpler two-category (background/foreground) partition. This improvement in critical areas is achieved by reducing the fidelity of less critical background elements, while the visual quality of other essential driving-related classes (weak ROI) is at least maintained.

In this paper, five classical edge detection operators are compared, and then a traffic signal light detection and recognition scheme that can be used for intelligent connected vehicles is implemented. Firstly, the image to be processed is obtained by detecting the traffic signal light through the vision sensor. The image is preprocessed: the color space of the image is converted from RGB space to HSV space. Through the grayscale, histogram equalization, image binarization processing, using the morphological closure operation, the five operators are compared in noise sensitivity, positioning accuracy and signal-to-noise ratio, the Canny edge detection operator is selected for image edge detection, and the target recognition area is obtained. Finally, using the histogram drawn, the number of red, green and yellow pixel points in the histogram can be clearly counted, and the color with the largest number of pixel points can be identified as the color of the identified traffic signal light, and the identification of the traffic signal light can be completed. The actual pictures are simulated on the MATLAB, which verifies the feasibility of the proposed method of traffic signal light recognition method based on Canny operator in this paper, which can correctly identify the color of the traffic signal light.

In the era of urbanization in the 21st century, the number of vehicles on the road has increased drastically to cope with the increased population of the city which directly affects the traffic on the street and more importantly on intersections. Automatic traffic signaling system can plays a vital role to mitigate vehicle congestion in the critical intersections of the busy roads. To ensure hassle free mobility of vehicles in the intersection on the street, the appropriate traffic signal lights are necessary that coordinate traffic in various directions on the streets. A simple intersection on the road is considered for the simulation of traffic signaling system. VHDL code was written in Quartus II to implement traffic light system for specific intersections on the road. The vehicles are always moving on the main road, meaning green light is always ‘1’ until there are vehicles approaching the side road. When the vehicles arrive on the side road which is sensed by sensor, then the traffic controller takes the initiative to schedule the traffic light, from green to red light in the main road and red to green light in the side road. The proposed traffic signaling system works properly, which is verified during simulation. In this paper, simple intersection is considered due to simple design and easy implementation.

Traffic light control and coordination is a critical function in today's busy roadways. Typical traffic lights have been shifting from fixed timing to ones that are based on a variety of sensors. Several shortcomings have been identified when considering these different approaches. In this paper, we propose smart dynamic traffic lights that can adapt their signaling times (e.g., changes from green to red and vice versa) according to the traffic density, exploiting Direction of Arrival and Timing Advance information transmitted from cell-phones, carried by the car drivers, to smart antennas installed on cellular base stations. The simulation analysis shows that the proposed system has the potential to reduce the queuing delay by 92%. Consequently, such an approach would reduce fuel consumption and pollution by 92% by avoiding queue on traffic signal lights.

As humanity has advanced, cars have become increasingly common on roads, and vehicle activity has skyrocketed. This shift has naturally created requirements for stop light systems to be implemented to control traffic flow through intersecting routes. This paper tackles the issues such as efficiency and pollution that arise due to the existence of these stop lights by exploring more efficient methods for a car to approach a stop light to save time and fuel. The different approaches are all basic logical methods tested in iterative simulations. I then found the best approach to travel from point A to point B when there is a stop light between the two points whose timing is dictated by a probability distribution.

This paper focuses on traffic sign spotting (TSS which automatically recognizes not only the conventional traffic signs but also information, facility and service signs, and traffic lights. TSS is divided into two sequential tasks: detecting traffic sign candidate regions in an image and recognizing the traffic signs in the regions. It is a very challenging task. We make the following contributions: 1) we create a traffic sign collection from the driverless car. The traffic signs are shot under the natural environment which covers large variation in illuminance and weather conditions. It not only contains the common traffic signs but also contains the information, facility and service signs which are called signposts, as well as traffic lights. 2) we proposed a systematic solution. We construct an Inception convolutional neural network. We use Faster-RCNN for traffic sign detection and make it suitable to detect small targets. 3) We adopt three schemes for the common traffic signs, the signposts and the traffic lights, respectively. The experimental results demonstrate the effectiveness and efficiency of our methods. Our methods won the first place in the traffic sign recognition task of Intelligent Vehicle Future Challenge 2017, China.

Intelligent transportation is a typical case of cyber-physical system (CPS). Due to the rapid increasing of the number of vehicles in city, problems caused by vehicles, like congestion and environment pollution, are becoming more and more serious. Traffic light control system is often used to control the vehicles passing for a solution of the congestion in the city. Present control systems used are normally assigned as to be static, i.e., traffic light signal changes in a static way. The aim of this paper is to propose a dynamical traffic light control system, i.e., change the traffic light signals in real time following the speed of vehicles. This system is an instance of V2I(Vehicle to Infrastructure) communication model, realizing data transmission between vehicles and traffic lights. Vehicles send speed messages to the traffic light when passing an intersection, then the traffic light analyzes the information and adjusts the signal time in real time. Each traffic light in each direction has a control strategy of itself without the orthogonal requirement. Therefore, the traffic light is a kind of cyber-physical system. This traffic light control system can maximize the number of vehicles passing intersection, and as a result, minimize the congestion and pollution. A traffic light control algorithm based on speed of vehicles and its simulation are presented. The safety and liveness of this control system are discussed too.

The walking speed of handicapped people generally is slower than that of normal people. So it is difficult for them to cross at crosswalks within the allotted time provided by the traffic light. This problem can be solved by expanding the time of the traffic light. However, if the latency of the traffic light is increased without distinguishing the handicapped among all other pedestrians, the efficiency of traffic signal lights will decrease. In this paper, we propose a smart traffic signal connecting mechanism between the previous pedestrian traffic signal and a pedestrian's device (smartphone). This Smart pedestrian traffic light, through this mechanism, minimizes traffic congestion by providing additional walking time only to the handicapped among pedestrians. This crosswalk traffic light recognizes the handicapped using a technique called Internet of things (IOT). In this paper, we extract the data necessary to build an effective smart crosswalk traffic light mechanism through simulation techniques. We have extracted different kinds of traffic signal times with our virtual simulation environment to verify the efficiency of the smart crosswalk pedestrian traffic light system. This approach can validate the effective delay time of the traffic signal time through a comparison based on number of pedestrians.

In recent years, with the utilization of blue Light Emitting Diode, the Light Emitting Diode element is beginning to be actively used in various scenes, and Light Emitting Diode is used even for traffic light now. The Light Emitting Diode traffic signal light has little power consumption as compared with the traffic signal light using the conventional electric bulb, and spreads to progress increasingly from now on. Because the discernment of the yellow of a signal light and red is difficult for sense-of-color unusual person, in the present signal light system, yellow signal light color is made somewhat brighter than red signal light. The aim of the present study is to collect the fundamental data for the display of the traffic signal light, which is easy to be recognized also by sense-of-color unusual persons. The results revealed that the optimal brightness of the yellow light from which a sense-of-color unusual person can distinguish the difference between yellow and red, i. e., “vision barrier-free domain”, is the area where yellow light is 2.8 times brightness of the red light.

Driving Assistance Systems (ADAS) have become integral to enhancing road safety and driver convenience in autonomous and semi-autonomous vehicles. One key component of ADAS is the accurate recognition of traffic signs and traffic lights to assist drivers in following road regulations and improving decision-making. This paper proposes a vehicular control system based on Computer Vision and Deep Learning techniques, designed to recognize traffic signs and lights and provide real-time driving assistance with distance estimation. The system employs Convolutional Neural Networks (CNNs) for traffic sign recognition and a hybrid detection model for traffic light detection, ensuring accuracy in dynamic environments. The proposed system was tested on publicly available datasets and demonstrated significant improvements in detection accuracy and response time, contributing to the development of safer and more efficient autonomous vehicles