The Andhra Pradesh government has granted administrative approval for Project Sarthi (Systematic Augmented Radial Traffic and Hoop Induction) in Visakhapatnam, with a budget of Rs 99.19 crore. This initiative aims to deploy an integrated AI-based intelligent traffic management system across 101 traffic junctions in the city.
Addressing Growing Traffic Challenges
With increasing traffic on major roads, beach corridors, BRTS routes, and NH-16 stretches, the government recognized the urgent need for a modern, real-time traffic management system. Under Project Sarthi, the Greater Visakhapatnam Municipal Corporation (GVMC) will introduce AI-powered traffic monitoring and enforcement systems. These include adaptive traffic signals, red-light violation detection, automatic number plate recognition (ANPR), helmet and triple-riding detection, wrong-way driving detection, and identification of non-standard number plates. The system will also support free-flow traffic monitoring, facial recognition at selected locations, pedestrian signals, and integration with police command-and-control centers.
Scope and Technical Specifications
The project will install adaptive traffic control systems at 101 junctions and violation detection systems at 90 junctions. Free-flow traffic monitoring will cover 14 locations, while facial recognition systems will be set up at 12 sites. Pedestrian safety signals will be established at five locations. Prior to approval, officials conducted a proof of concept for the facial recognition component at Satyam Junction in Vizag in March. The system was tested for identifying individuals wearing helmets, masks, caps, and goggles, as well as matching older photographs with current images using AI technology. Officials reported that the successful trial bolstered confidence in the system's utility for policing, public safety, and enforcement.
Adaptive Traffic Control and Green Corridors
A key feature of Sarthi is the adaptive traffic control system, which automatically adjusts signal timings based on real-time vehicle movement. Officials stated that this will reduce waiting times, improve traffic flow, and ease congestion at junctions. The system will also facilitate green corridors for emergency vehicles and VIP movement on important routes, including NH corridors, railway station road, and BRTS stretches.
Ministerial Remarks
State Municipal Administration Minister P Narayana described Project Sarthi as a major step toward making Visakhapatnam a safer and smarter city. "The project will improve traffic discipline, reduce congestion, enhance road safety, and support citizens with smoother and more predictable movement across the city," Narayana said.
Official Perspectives on Sustainability
Principal Secretary of Municipal Administration and Urban Development S Suresh Kumar emphasized that Sarthi is not merely a traffic signal improvement project but a larger AI-enabled urban mobility platform. "The project has been structured to ensure not only installation of technology but also long-term functionality, accountability, and sustainability. The escrow mechanism and performance-based payment structure will ensure that the system continues to operate effectively over the five-year operational and maintenance period," Kumar explained.
Financial Structure and Implementation
- The government approved 60% capital expenditure (capex) support worth Rs 49.54 crore for the project.
- GVMC will bear the remaining capex and five-year operation and maintenance cost of Rs 49.65 crore.
- The amount will be paid bi-annually using GVMC funds and revenue from e-challans and enforcement collections.
- The government also allowed VMRDA to contribute Rs 20 crore toward the project.
- A dedicated escrow account will be maintained by the GVMC commissioner for Project Sarthi, used exclusively for implementation and operation.
- Revenue from e-challans, traffic penalties, and enforcement collections will be deposited into the escrow account.
- Payments to the selected agency will be released after verification of milestones and performance.
