Traffic jams are a long-standing transportation issue that has plagued us for decades in Bangalore. It is an ever-increasing threat to practically all of India’s major “IT centres,” with Bengaluru being the worst affected. Once a paradise of lakes, gardens, and pleasant weather, Bengaluru is today a metropolis of overcrowded automobiles emitting toxic fumes. As per Metrosaga, two-wheelers continue to rule the roadways, accounting for up to 73% of the city’s total vehicle population, followed by cars (20%). While the times have changed, the solutions to the traffic management problem in most Indian metropolises have not. Urban traffic will be revolutionised by artificial intelligence (AI), which will eliminate bottlenecks and chokepoints. As a result, we will be able to minimise both traffic and travel time. The effort is expected to address Bengaluru’s traffic problem.
How can AI help with Bengaluru traffic?
With cities like Bengaluru battling traffic congestion, artificial intelligence (AI) solutions for receiving real-time data from vehicles for traffic management, and exploiting mobility-on-demand in trip planning through a single user interface, have emerged. Other options for effective traffic management include the safe integration of AI-based decision-making, traffic management, routing, transportation network services, and other mobility optimisation tools (Transportation, 2019).
Sign up for your weekly dose of what's up in emerging technology.
Several of India’s public services are slowly adopting artificial intelligence (AI). As per the report all traffic signals in Bengaluru would be redesigned with the help of artificial intelligence (AI), a long time ago. The Bengaluru Police Department is the latest to join the club, with plans to install AI-based traffic signals at most city intersections. Intelligent Transportation Systems (ITS) have been developed and deployed during the last three decades as a cost-effective investment to boost the capacity of existing transportation systems. Accurate assessments of real-time traffic operational circumstances are crucial for the efficient administration of ITS.
Contributions to the Field
Over the last three decades, ITS technologies and solutions have been developed and deployed as a cost-effective way to increase the capacity of current transportation systems. According to S. H. Shah et al., special sensors (such as magnetic coil and ultrasonic detection) are routinely put along the roadside to monitor traffic conditions. However, the deployment of these sensors necessitates changes to the traffic infrastructure. In addition, due to transmission difficulties, their sensory data is most likely limited in quantity and quality.
Agachai Sumalee et al. stated that a video-based traffic counting system (VTCS), compared to a sensory-based method, simplifies deployment and maintenance operations and seamlessly interacts with AI to detect, track, and identify each vehicle and its activities. Artificial intelligence has made some notable contributions to a vehicle traffic study. Some of them are listed below.
- The regression-based methods detect and count the number of cars by utilising global image features (e.g., colour histogram and pixel density).
- For density estimation, V. Lempitsky et al. offer the Maximum Excess over SubArrays (MESA) distance concept.
- Mingpei Liang et al. introduced a cascade regression approach to assess and classify vehicles based on low-level data for each foreground segment.
- To regress vehicle spatial density photos, H. Tayara et al. uses a convolution neural network (CNN).
- Meanwhile, Zhang et al. count the number of cars by extracting spatial-temporal information from density maps using an FCN-rLSTM network.
- A. Dame et al. offer the Kanade-Lucas-Tomasi (KLT) feature tracker based on pyramid representation.
- C. Choi et al. suggest spatiotemporal relations that can be used in a video to extract vehicle movement trajectory data.
- A. Bewley et al. combine the Kalman Filter and the Hungarian algorithm to create Simple Online and Real-time Tracking (SORT), which supports multi-object tracking while maintaining good performance.
- From counting vehicle results, Y. Wang et al. determine a traffic stream’s average speed, density, and volume.
- Gao et al. classified vehicle sizes as lite, small, medium, and large based on vehicle motion flow.
Earlier, NVIDIA with IEEE and academia conducted the first AI City Challenge to stimulate the development of efficient algorithms that have direct benefits on smart transportation. Moreover, the CVPR 2021 AI City Challenge workshop is concentrating on ITS issues.
This article is merely an investigation into Bengaluru traffic and how AI can help. Recent research states that artificial intelligence raises a slew of ethical, societal, economic, and legal concerns. In addition, because AI-based systems are data-driven, they face cybersecurity and data privacy concerns, particularly in autonomous vehicles. However, there are a variety of AI-based traffic solutions. This article aims to identify and collect numerous AI strategies that can be used to solve specific difficulties in Bengaluru traffic. For Bengaluru traffic, we can certainly expect appropriate solutions from Indian researchers and startups.