Pixuate's One of a Kind AI-Enabled Traffic Survey, Karnataka

India’s 5.89 million kilometers (km) road network is the second largest in the world and it is our cardiovascular system—a multifaceted structure that enables the continuous flow of 90% of people and 64.5% of goods through its arteries. Our Government and its agencies are well aware that its quality and efficiency are crucial for our economy and the well-being of citizens.
 

To make informed decisions the governments’ road departments conduct regular traffic surveys to help analyze the usage of roads and help in better design of the roads that pass through villages, towns, and cities. At present, most traffic-related surveys for collecting data to aid in the planning and analysis is obtained through randomly sampled manual surveys over 48 hrs time period with the aid of sensors or going through gigabytes of CCTV footage manually which are as well highly inefficient mechanisms.
 

Pixuate by Cocoslabs Innovative Solutions Pvt. Ltd is honored to have partnered with Public Works Department(PWD), Government of Karnataka to conduct state-of-the-art AI-powered traffic surveys at 110 locations in Sirsi Division. This is one of the very first instances of using AI at scale for traffic surveys.
 

Pixuate’s traffic survey has enabled the Public Works Department, Govt. of Karnataka to have insights into the Average Daily Traffic in the Sirsi Division which will play a key decision-making role while deciding upon the design of pavements, improvements to be made to the existing roads, whether to upgrade roads to higher categories or is there a need for widening some of the roads, etc.
 

Why adopt AI for traffic surveys?

Data accuracy of traffic surveys is of utmost importance, as any developmental activity on the roads is done based on the data made available via reports, and these developmental activities are funded by taxpayer money, so it’s important that every single rupee is put towards its best utilization.
 

Until recent times PWDs across the nation have been conducting traffic surveys manually, this has led to several instances of inaccurate data. The PWD, Govt of Karnataka is one of the first to have adopted an AI-based traffic survey for Sirsi in partnership with Pixuate, a leader in the Vehicle counting & classification space & the winner of SCOUT 2021 (SmartCity Challenge & Outcome for Urban Transformation) for its Automatic Traffic Count and Classifier (ATCC) platform with ANPR capability. The AI-enabled surveys are set to drastically bring down the data inaccuracies and enable data-based decision-making for Government agencies.
 

How did Pixuate overcome the challenges in conducting the traffic survey in the remotest regions?
 

Along with using its AI platform for this survey, Pixuate also deployed an advanced hardware setup to reliably capture traffic data.  In order to ensure that this first of its kind AI-enabled traffic survey of such a huge scale spanning 110 locations spread across an area of 200 km – 300 km road network in Sirsi, Siddapur, Yallapur, Mundgod, Haliyal & Joida taluks of Karnataka is successfully completed in a record span of 48 hours, Pixuate put together a lean-agile expert team that conducted extensive on-ground preparation & coordination with various stakeholders for over 7 days to prepare the region for the survey.
 

Key highlights that went into planning and executing a traffic survey of scale in the remotest parts of Karnataka
 

1. The locations which did not have electricity supply were installed with sources of alternate power supplies (Solar panels & Power banks) as part of the pre-survey work.
 

2. In order to not take any chance, the Pixuate project team identified locations with likely power supply failure and provided them with power backups as these areas had frequent power cuts.
 

3. A separate location team was put in place to do this extensive fieldwork of identifying areas with power supply issues and put in place power backups, along with mounting cameras for the actual survey to be conducted.
 

5. A QA team was tasked with regular monitoring of on-field camera recordings and making sure that all assets are intact.
 

6. To communicate at scale WhatsApp groups were created (location-wise) and the status of the surveys was shared real-time basis with the Engineers of each sub-division. In this way, any abnormalities could be sorted immediately.
 

The geographic challenges were as well high for this survey as several areas involved in the survey were part of forest regions, and a few regions were part of the Tiger reserve belt of Joida & Haliyal, the distance between one location to another was as well high coupled with very poor cellular connectivity which made this one of the challenging surveys to conduct a first of its kind AI-enabled traffic survey.
 

Overcoming all the odds
 

The successful completion of the survey in record time further testifies that Pixuate is the true, made-for-Indian roads platform solution with no hardware dependency and minimal human intervention. As with a lean team, Pixuate was able to pull off the survey.
 

Today, the PWD, Govt of Karnataka is in possession of accurate traffic data reports of the Sirsi Division, with image and video proof for verification, with a comprehensive AI-generated report of class-wise vehicle count for each of the 110 locations.
 

In this survey, Pixuate has reduced human errors, training overheads, and the time involved in traffic survey report generation. Pixuate’s platform is designed for scalability & capable of generating reports with high accuracy on a near real-time basis.
 

Pixuate’s Automatic Traffic Count & Classifier System
 

To automate the traffic survey process which needs high accuracy and faster turnaround time, Pixuate offers a complete system for traffic survey execution.
 

The hardware component of the system is the camera with power backup and data connectivity which can be easily installed at even remote locations of the survey. The traffic data captured by the camera is processed with one of the industry’s best turnaround time with Pixuate’s proprietary AI platform for traffic count and classification, built for Indian roads and vehicle types. Pixuate’s software delivers 98%+ accuracy and can be availed as a service for any traffic survey by agencies and government bodies across the globe.
 

At Pixuate we are also rethinking Vehicle Speed Detection Systems, read here to know more. We believe that this is just the beginning and Pixuate will play a major role when India opens up for free-flow tolling!