TrafficInfraTech 2017: Smart Mobility in India

We have just returned from a week in Hyderabad, attending the TrafficIntraTech 2017 exhibition. Attending every show since its inception, we have been excited by the changes taking place in India. In the words of our Business Development Manager, Maurice Berger, “TrafficIntraTech is a great avenue to meet with clients, explain what we do, find out what they do, answer questions and offer support”.

The TrafficInfraTech 2017 conference

With a focus on Smart Mobility, this year’s high-profile speakers discussed new technologies for traffic data collection, management and analysis, and sustainable strategies for transport development. The event’s location, Hyderabad, was timely given the city is a front-runner in adopting Intelligent Transport Systems. India is undergoing rapid change, with an ambitious plan to transition vehicle sales to all-electric models by 2030 and increase its shared mobility scheme.

The entire first afternoon was dedicated to traffic data. Our friends from DataCorp Traffic shared insights into collection methods and technologies. As long-term users of our RoadPod® VT tube counter, they discussed the importance of accurate axle data. This session tied in perfectly with Dr. V. Ravinder’s (Hyderabad City Traffic Police) discussion on the challenges of bad data.

It was great to catch up with our clients and we are looking forward to next years show.

MetroCount in India

For the past 10 years, we have collaborated with local authorities, survey companies and engineers to develop a vehicle classification scheme that is accurate and relevant to Indian roads. This has proven quite challenging for various reasons. Firstly, poor lane discipline across the country leads to outliers in the data, meaning some vehicles are not clearly identified. Also, similar axle and wheelbase configurations of common Indian vehicles make them difficult to distinguish.

While video monitoring systems are widely used to classify traffic, their accuracy and costs can make them prohibitive. In 2016, a large scale classification study emphasised the downfalls of video data collection in dense, urban traffic. Surveillance quality issues made data collection and validation challenging and time-consuming, while the resulting data was still unsatisfactory due to a lack of a better classification scheme. For this reason, MetroCount has invested resources in developing a more precise and suitable scheme for classifying vehicles in India.

Assumptions for video classification as presented in the “A large-scale dataset for classification of vehicles in urban traffic scenes” paper.

Misclassification example as presented in the “A large-scale dataset for classification of vehicles in urban traffic scenes” paper. From left to right: Heavy Vehicle classified under Light, Light Vehicle classified under Heavy, Merged vehicles classified under Heavy.

TrafficintraTech 2017 – Photo Gallery

Leave a Reply

Your email address will not be published. Required fields are marked *

To improve your browsing experience, this website uses cookies.

Cookies are small text files that enable us to record your website usage in accordance with our privacy policy.
Please confirm if you accept our tracking cookies.