How can an organisation leverage a network of smart infrastructure in the built environment to make better and more timely sense of emergency incidents such as building collapses and road traffic accidents, and to trigger early intervention measures, without the need to activate precious emergency resources?
A team of SMU students from the School of Computing and Information Systems - Patrick Lim Yan Hong, Vinnie Chu Yi Xuan, Ong Yan Kai, Jane Seah Hing Kid, and Lim Pei Xuan - took up this challenge and beat more than 370 teams in Singapore to clinch the first position at the SCDF x IBM Lifesavers' Innovation Challenge: Call For Code 2020. They were awarded with a cash prize of S$10,000 and an additional USD$120,000 in IBM Cloud credits to implement their solution.
The 48-hour hackathon was held virtually on 12-14 June, with the Grand Finals and Award Presentation Ceremony held on 18 July at the Singapore Civil Defence Force (SCDF) Headquarters. The teams were assessed primarily on their creativity, analysis of problem and implementation of strategies.
The team’s solution, called Dr Watson To The Rescue, utilises smart infrastructure such as sensors embedded within buildings to increase the efficiency and safety of emergency evacuation and rescue. The system can be deployed to any building as long as they have the infrastructure required, including WiFi, CCTV cameras and smoke/fire sensors.
Dr Watson To The Rescue comprises of four modules:
- Route Finder: helps building occupants and responders find the shortest escape and rescue routes;
- People Counter: helps rescuers find where occupants are congregated in the building;
- Mobility Aid Detection: helps rescuers locate where at-risk individuals such as mobility-impaired occupants are in the building;
- Distress Identification: a contingency module which recognises screams for help when CCTVs can no longer be relied upon due to smoke
During a fire outbreak in a building, the lack of real-time information for rescuers and building occupants have often brought about delayed evacuation and even casualties due to the occupants running to blocked exits and obstacles, building security not knowing where urgent assistance is required, and the SCDF not having real-time information of stranded occupants and state of damage to the building structure.
For building occupants, the route finder module in Dr Watson To The Rescue uses WiFi Triangulation and algorithms to locate occupants within a building and provide personalised escape routes to their mobile devices. These routes are optimised in real-time, taking into consideration factors such as obstruction, congestion, as well as location of the fire.
For responders, upon trigger of the fire alarm, the CCTV footage from the site will be processed to automatically identify crowded zones, as well as areas where individuals with mobility impairments were detected (e.g. wheelchair bound occupants). Using Artificial Intelligence algorithms on sounds captured by sensors, the distress identification system will also locate stranded victims calling for help, thus providing on-the-ground information even when smoke obscures the CCTVs. The security team in the building will be given an optimal and safe route based on Dr Watson To The Rescue’s route finder module to get to these critical areas to provide early assistance before SCDF arrives at the scene. At the same time, the SCDF will also be fed with real-time updates via Dr Watson To The Rescue’s dashboard to strategise efficiently and effectively as they rush to the scene.
Team leader Ong Yan Kai said, “The team split the work such that the dashboard and each module is developed by one person. We spent about one day developing the individual components, before coming together to integrate them.”
“Working remotely as a team required a high level of mutual trust and discipline. As all of us were either working or attending postgraduate classes, we did not have the luxury of time to schedule multiple meetings or check on each other’s progress. Hence, it paid to have team members who possess complementary skills and who share the same drive and determination to succeed. Such a team is like a well-oiled machine. Despite the limited virtual meetups, we hit the ground running at every meeting and brainstorming session. The team achieved synergy relatively quickly because each team member was consistently putting his/her best foot forward. When you have a crew of multipliers rather than mere additions, creativity and productivity soars. Undoubtedly, I would say this was the key to our success in this competition. I am truly grateful for such an amazing team.”
The SMU team developed their solution using technologies such as React (web dashboard) and Flask (Route Finder web service), hosted on Amazon Web Services’ Simple Storage Service and IBM Cloud Foundry, respectively. IBM Watson Visual Recognition was used for object detection (for the Mobility Aid Detection and People Counter modules), while the Distress Identifier was based on IBM Developer Model Asset Exchange: Audio Classifier (for the Distress Identification module), as well as the Internet of Things.
Dr Watson To The Rescue is currently a proof of concept, with each module in the system being successfully implemented on a small scale. The SMU team will be working closely with SCDF to conduct further knowledge sharing sessions with the relevant stakeholders.
The "SCDF x IBM Lifesavers' Innovation Challenge: Call For Code 2020" was organized with the aim to spread awareness and involve more Singaporeans in #TechForGood, by ideating and co-creating applications using cutting-edge technology to make Singapore a better and safer place together.
This year, a record number of more than 370 teams comprising 1244 students from junior colleges, polytechnics and universities competed in the challenge.
Caption: SMU student Ong Yan Kai received the winners’ cheque of $10,000 from SCDF Deputy Commissioner (Future Technology & Public Safety) Teong How Hwa. To his right is Ms Natasha Kwan, General Manager of Enterprise Public Sector, IBM Singapore. [Back row, L-R] Patrick Lim Yan Hong, Vinnie Chu Yi Xuan, Jane Seah Hing Kid, and Lim Pei Xuan.
Last updated on 04 Aug 2021 .