Most of us who call Dhaka home have a love-hate relationship with the city. For decades, a relentless challenge has cast a shadow over the city’s vibrant spirit: Traffic congestion. This enduring issue has not only tested the patience of its residents but also hindered the smooth flow of life. Despite various stakeholders’ countless discussions and efforts to address this predicament, tangible solutions have remained elusive, much like the sluggish crawl of vehicles on the city’s roads.
The city’s unrelenting traffic jams, characterized by their excruciatingly slow pace and seemingly interminable gridlocks, exact a staggering toll on the local economy. According to a December 2021 report from a local daily The Business Standard, traffic congestion in Dhaka cost the nation 2.9% of its GDP. In monetary value, this equates to a staggering BDT 1010.36 billion. Daily, an average commuter loses 3 to 5 hours of productive time navigating the congested Dhaka streets. The fuel wastage resulting from idling vehicles and stop-and-go traffic costs Dhaka residents an additional USD 1.5 billion annually, diverting resources from productive investments.
The traffic congestion in Dhaka is rooted in a complex interplay of multiple factors. Professor Dr. Md. Hadiuzzaman, ex-Director of Accident Research Institute, BUET, has identified weak traffic management and inadequate infrastructure as the biggest contributing factor towards the deadly Dhaka traffic.
Continuous efforts have been made to formulate and implement new transportation policies. Despite all the trials and errors in traffic management and all the new flyovers and U-loops forming a concrete maze over the city – the problem persists and is stronger than ever. Could technology perhaps come to the rescue?
Over the past two decades, approximately BDT 2 billion has been allocated to implement advanced technologies such as signal lights, traffic countdown timers, and the ambitious Intelligent Traffic System (ITS). However, according to Implementation Monitoring and Evaluation Division (IMED) reports, the efficacy of these measures remains questionable, with hardly any intersection experiencing notable improvements—save for the exception of the Gulshan 2 signal. In 2018, DTCA (Dhaka Transport Coordination Authority) launched an ITS (Intelligent Traffic System) on an experimental basis in Mohakhali, Gulshan 1 and 2, Paltan, and Fulbaria. Although the project cost us BDT 0.52 billion, no positive result emerged.
From writing thesis proposals to creating stunning images – suddenly, it feels like AI or artificial intelligence is a part of every conversation. How can this technical innovation come to the rescue of Dhaka dwellers?
Amidst the ongoing experimentation with various strategies and initiatives, a notable reservoir of support beckons from modern AI technologies, encompassing the prowess of Computer Vision, the Internet of Things (IoT), and Data Mining. A hidden potential lies within the sprawling network of CCTV cameras stationed across diverse locations within Dhaka city, predominantly utilized for surveillance purposes. However, these unassuming cameras could transcend their traditional roles, transforming into potent tools for vehicle quantification at signal points.
The city boasts an extensive fleet of CCTV cameras, each a silent observer capturing the city’s pulse. These cameras dutifully relay live video feeds through dedicated network bandwidth to various authoritative repositories known as Network Video Recorders (NVRs). This video reservoir empowers authorities to retrospectively review the footage, extracting insights and discerning trends from the ebb and flow of Dhaka’s traffic patterns. Yet, this reservoir of video feeds remains untapped in its potential to provide real-time analyses and meaningful insights.
This is where the image-processing software comes into play. With remarkable finesse, such software could transform the live stream of video data into a dynamic source of real-time analyses, unraveling an array of invaluable features from the camera feeds. Notably, this includes discerning vehicle types, meticulously tallying vehicular numbers, pinpointing instances of lane violations, and even reading license plates with remarkable accuracy. The synergy of AI-powered image processing and the wealth of available video data could instigate a transformative leap in how we perceive, manage, and optimize traffic within Dhaka’s urban landscape.
As we venture deeper into the capabilities of AI-driven technologies, the potential to revolutionize traffic management by leveraging these tools becomes increasingly tangible. This article will chart a course through the diverse dimensions of AI applications, unearthing possibilities that may redefine the contours of Dhaka’s traffic narrative.
If we think about the sheer volume of the available video data through CCTVs – it is evident that the analysis demands a significant investment of working hours. However, the true need lies in extracting meaningful information, or metadata, from these videos.
Sigmind.ai, a Dhaka-based startup specializing in computer vision, has undertaken noteworthy collaborations with various authorities. CEO Abu Anas Ibn Samad explained that their technology can efficiently tally vehicles on the road and relay real-time reports on vehicle types from streaming videos. Their software can be deployed on the network’s immediate ‘edge’ and central servers.
Sigmind.ai’s contributions extend to working closely with the Law-and-Order Coordination Committee (LOCC), a cooperative platform involving locals, police, DNCC, and community organizations.
Notably, the cameras stationed at entry and exit points of the diplomatic zone can adeptly read Bangla license plates. While LOCC’s primary goal is to uphold safety in the diplomatic zone and adjoining areas, the AI-powered software has been harnessed to extract data that enhances comprehension of vehicle movement patterns and types within the area. This has facilitated a deeper insight into traffic volume, enabling more effective route management.
Anas asserts that the considerable number of CCTV cameras strategically positioned by various government bodies can be repurposed for traffic data analysis with minor adjustments. Their technology further excels in detecting assorted anomalies caused by both vehicular traffic and pedestrians. The processing of traffic video data can occur either directly at signal points using their embedded devices or remotely at a central server. Ultimately, this data can be presented to users through a centralized dashboard, offering a comprehensive understanding of traffic congestion dynamics.
Adding to the roster of challenges in Dhaka’s traffic landscape is the inadequate parking infrastructure, further exacerbated by the prevailing disorderly parking practices. However, even within this conundrum, noteworthy developments have emerged to tackle this issue head-on. Companies such as ParkingKoi and ParkKori have stepped into the arena, offering real-time, cloud-based parking management solutions. Their comprehensive platforms encompass features like live parking space availability, guided navigation to available spots, and seamless online reservations accompanied by real-time status updates.
In the intricate tapestry of Dhaka’s traffic challenges, AI emerges as a pivotal thread, weaving promise into the city’s mobility landscape. However, the solution to this multifaceted issue lies not in isolated endeavors but in the symphony of collaborative efforts. While not an instant fix, AI can amplify the impact of coordinated actions across decision-making and implementation levels. This transformation, bolstered by the ingenuity of local techpreneurs, can reshape Dhaka’s traffic dynamics and lead our city towards a more efficient tomorrow.
Hasib Bin Rafique a Software Developer and Designer, with a special knack for computer vision software.