AI may be used to detect accidents in real time in Malaysia

November 18, 2024 - 09:23
Dr Law, who is the head of the Road Safety Research Centre at Universiti Putra Malaysia’s Engineering Faculty, said Arids – which was introduced in February – is in pilot phase at portions totalling 1,000km at expressways and federal roads in Klang Valley and Kuching.

 

A three-vehicle pile-up along Jalan Bernam, Malaysia in December 2023. THE STAR/ANN Photo

PETALING JAYA -- The Malaysian Highway Authority (LLM) is mulling over the full adoption of the Automatic Road Incident Detection System (Arids), an artificial intelligence-based solution to detect road accidents in real time to reduce rescue response times on expressways.

Road safety expert Assoc Prof Dr Law Teik Hua, said Arids, which is currently not used to its full potential, can facilitate quicker responses from first-responders to accidents.

Dr Law, who is the head of the Road Safety Research Centre at Universiti Putra Malaysia’s Engineering Faculty, said Arids – which was introduced in February – is in pilot phase at portions totalling 1,000km at expressways and federal roads in Klang Valley and Kuching.

“However, Brunei and the provincial capital city of Xi’an in China, have adopted the Malaysian-made system in their bid to make their roads safer.

“We have presented to LLM on the need to adopt it fully throughout Malaysia,” said Law.

When contacted, LLM director-general Datuk Sazali Harun confirmed that it is seriously considering its full implementation.

Dr Law, whose team developed Arids, said the fatal crash involving the container lorry in Simpang Renggam, Johor, on Nov 15 was detected by Arids at 7.46am that day.

“However, checks showed that the earliest official report from the North-South Expressway (Plus) Trafik X channel stated that the incident occurred at 8:09am.

“If the crash was detected earlier, rescue response time would have been cut shorter, with better survival rates.

“Most accidents on highways in Malaysia are attended to based on CCTV monitoring by the respective highway concessionaire, or reports from users or highway patrols,” said Law.

Currently the authorities in Xi’an are using it on roads covering 200 traffic lights in a pilot run.

“Brunei has taken it and implemented it for its capital city to fully monitor traffic light junctions and also detect accidents and abnormalities,” said Dr Law, who is also a consultant with LLM, Kuala Lumpur City Hall (DBKL), and the Federal Traffic Investigation and Enforcement Department (JSPT) of the police.

With the core of the system based in Malaysia as a mobile traffic management system that allows remote access, Arids can be accessed using a mobile phone, and runs continuously using Google traffic data to provide accident notifications as well as analyse traffic congestion data, including analysing whether traffic lights are functioning.

“The system detects accidents autonomously in real time as well as vehicle breakdowns and other traffic situations utilising neural networks (a type of machine learning process that is inspired by how humans learn),” said Dr Law, adding that Arids users can be automatically notified of anomalies via WhatsApp, without the need of any human input.

He said analysis of crashes on the system also showed the need for sturdier guardrails on highways, which currently can only withstand the impact from smaller vehicle crashes.

In response, West Coast Expressway (WCE) CEO Lyndon Alfred Felix said while crash prediction systems are continuously evolving, WCE has already implemented technology-driven solutions such as Traffic Monitoring Systems (TMS) and CCTVs.

“These systems allow us to identify potential risks and patterns – such as erratic driving behaviour or excessive speed. WCE is also exploring advancements in predictive technologies,” he said, adding that the extent of enforcement is often dependent on coordination with relevant authorities.

“The implementation of a Weigh-In-Motion (WIM) system will enable real-time monitoring of vehicle loads without requiring them to stop, thereby allowing the concessionaires to address issues related to overloading proactively,” said Felix.

Quicker accident detection, naturally, needs to be paired with more stringent inspections on heavy vehicles, especially by concessionaires at the entry points of closed-toll networks.

Dr Law said current laws do not permit concessionaires to do so, hence the reliance on enforcement authorities.

Sazali concurred, saying concessionaires currently have no legal authority to stop overladen lorries or unsafe vehicles from entering their networks.

“All vehicles are allowed to use highways or expressways except those restricted via restriction orders gazetted under the Road Transport Act, which also places WIM under the jurisdiction of the Transport Ministry,” he said. -- THE STAR/ANN

E-paper