Researchers say a new dispatch system uses artificial intelligence to improve taxi utilisation and passenger waiting times.

The smart dispatch decision system is jointly developed by the faculty of engineering at the University of Hong Kong and the SynCab taxi fleet.

The system, in use since the launch of SynCab's fleet in early July, currently covers more than 160 cabs.

The director of the Smart Mobility Lab at the university, Professor Ke Jintao, said the system can, compared to the broadcasting system used before, better guarantee that drivers who are closest or most suitable get to pick up customers.

"We use AI algorithms to consider multiple factors, including weather, traffic congestion ... to determine the best order for each driver."

Ke says the system is better suited to electric cars with its tailored algorithm as it uses machine-learning models to predict remaining battery power.

The system, he added, will continue to improve as more data comes in with more drivers and passengers.

"Then we will retrain the data and have a more efficient version of the algorithm, improve further, and we can have a win-win situation again."