logo
POST TIME: 12 January, 2019 12:26:39 AM / LAST MODIFIED: 12 January, 2019 12:09:32 PM
Artificial intelligence-based road monitoring on cards
FAISAL MAHMUD, Dhaka

Artificial intelligence-based road monitoring on cards

The ICT Division plans to install a road monitoring system based on artificial intelligence (AI) to check the northward trend of road accidents and ever-increasing traffic congestion in the country’s roads and highways. Sources said the plan for installing AI-based devices along the country’s highways has been placed in the ICT Division’s “to-do list” for 2019. The ICT Division has already appointed a local company, eGeneration Ltd, for laying out the design of such AI devices. Talking with The Independent, eGeneration Group chairman Shamim Ahsan said that the root cause of the dual problems of road accidents and road congestion should be detected to devise an effective solution. Use of AI for drivers can dramatically reduce the number of road accidents, he added.

Shamim Ahsan also said that most traffic incidents around the globe were predictable and preventable. Use of AI or any smart device could help the authorities in this regard, he added. “About 90 per cent of accidents occur due to human error. Modern technology like automated emergency braking, lane departure prevention and autonomous driving will be important for vehicle safety,” said Ahsan.

The eGeneration Group chairman also said that eGeneration Ltd was consulting Bangladeshi expatriates working in other IT companies in the US to zero in on cost-effective and implementable deep learning-based monitoring devices for roads.

Shehzad Noor Taus, a young Bangladeshi engineer from the Deep Learning Division of the world-renowned NVIDIA in the US, is already working with eGeneration Ltd to work out a possible solution for designing an intelligent device to monitor roads.

Taus, who came to Bangladesh recently, said AI could address the issue of safety in different ways. “An intelligent system can address the issue through the advanced driver assistance system (ADAS). There are solutions like driver drowsiness detection, where the AI tool can detect if the driver falls asleep during a drive. This tool would be able to alert the driver,” he added.

Taus also said drivers had to adjust their speed to desired levels to accommodate the needs of other road users.

Another technology called dynamic speed adaptation (DSA) offers remote speed advice to drivers about appropriate speeds in particular stretches of roads. It has already been tried in Sweden and the Netherlands.

“We have seen that cameras can measure the speed of a vehicle over long distances. CCTV cameras may be installed on all traffic signals where they may read the number plates of each and every speeding vehicle and instantly compare them with a national database. This would help law enforcers intercept illegal vehicles,” said Taus.

The young Bangladeshi engineer said there was a lot of interest in autonomous driving. Significant investments have been made by companies and governments to conduct research on this area. “But we are several years away from full autonomy and need to invest now in our drivers to improve road safety,” he added.

He, however, said that much of the technology and learning from the area of autonomous driving could be transferred to improve the manned vehicle experience. High-definition cameras and onboard sensors can be used to capture the environment around the driver and the vehicle, he added.

This information is analysed in real time using powerful integrated processors and deep learning algorithms. Any unsafe event or incidents of concern are sent to the monitoring centre along with the relevant video and sensor information.

“This includes intelligent detection of events such as traffic light violation, lane departure, tailgating, sign violations and relative speeding, hard acceleration and hard turns, among others. These solutions help the monitoring centre to watch the vehicles closely and react in real time. It reduces the possibility of risky driving becoming a critical event,” said Taus.

He also said there were some parameters to measure and manage every driver’s performance on the road. This would help the authorities to differentiate between good and bad drivers, he added.

MK