• 14 November 2024
  • Challenges

  • by ,

TUMI Lab HCMC: Big Data for Climate Proofing Urban Buses to Flooding

by Assoc. Prof. Vu Anh Tuan, dated 27.09.2024

 

Current achievements

The TUMI Lab Ho Chi Minh City (HCMC)  (2022-2024) is been implemented  by the Vietnamese-German University under the GIZ, Transforming Urban Mobility Initiative funded by the German Federal Ministry for Economic Cooperation and Development (BMZ).

The main objective of the TUMI Lab  is the design and development of a  digital tool (so called Smart Move System – SMS) that utilizes GPS bigdata and machine learning algorithms for automating short-term predictions of traffic congestion and the rerouting of affected bus lines under the influence of extreme rainfall and flooding events in Ho Chi Minh City (HCMC). The SMS provides user-interactive functions through which predicted congestions and bus route rerouting can be communicated to bus operators for implementation and bus passengers or road user for travel adjustments. An outstanding function of the SMS tool and what sets it apart from Google Maps, is the capability to forecast the state of traffic in the next 30 minutes to 1 to 3 hours that Google Maps cannot offer. This function is extraordinarily important for dynamic urban traffic management in response to ever-changing climate events. The Lab is the first attempt in Vietnam to deploy advanced ML models to forecast average bus speeds on a real-world bus network.

Road flooding and resulted road congestion in Ho Chi Minh City

 

Approach to Short-term Traffic Speed Prediction

 

Short-term traffic congestion prediction (app)

 

Necessary improvements for the success of the model

The LAB is now working to improve the AI models in a few areas to make it more powerful and ready for real-world deployment. First, multi-year GPS data needs to be complied for improving the ML models in terms of pattern learning and prediction. Second, complex AI models should be developed to deal with the issues of discontinued data points and limited datasets for model training. Third, the modeling of bus route networks and urban road networks must be enriched to allow for more accurate estimates of average bus speeds on the network.

The role of an application server

VGU is going to install an application server at the university campus to support the necessary improvements. Professors, researchers and students of the university will continue working to enhance the SMS tool. HCMC Department of Transport (DOT) promises to continue their support through data provision and system piloting in the coming year. VGU is also seeking further funding to support R&D activities in relation to the needed improvements. All of these efforts together will contribute to the scalability and long-term sustainability of the SMS tool.  The LAB hopes to extend and contextualize the SMS tool to other Vietnamese and Southeast Asian cities.

Expected impacts

The complete SMS tool enables prediction of early traffic congestions, which are jointly caused by recurrent traffic and climate events (e.g., heavy rainfalls, high tidal rise, and flooding). Having received the information, bus operators will timely adjust dispatch plans, avoiding gridlock and sending buses to high congestion points. This will help ensure continued and smooth bus operations which benefit both riders and local transport authorities and their workers.  Further,   local government  will improve their capability in monitoring, forecasting and responding to  ever-increasing climate events and associated traffic chaos, thereby enhancing climate resiliency.

Related

Contact