Main Page

SMT483: Project Experience

Welcome to this course wiki for SMT483: Project Experience.

You can access the Project Groups page here, where you will write your group projects.

Please note that:

• This wiki is available for anyone in the world to view, please do therefore not post any personal information on this wiki.
• You need to be logged in with your SMU username/password to edit the content.
• You can read the help pages and view this video tutorial to learn how to use the wiki.
• Please make sure that you do not violate Intellectual Property Law's. You will find a guide here, which will help you to determine if your content is fine to post. In this document you will also learn how you can find and post photos (from Internet) legally on this wiki.

Course Information

Faculty	TAN Hwee Pink
Course	Project Experience
Course code	SMT483
Term	AY2019/2020 Term 2
Section	G1
Teaching Assistant	Vinnie Chu

Potential Projects

Number	Sponsor	Project Topic	Project Description and Deliverable	Skill Requirements
1	URA	Modelling vehicular flows into Pick-Up Drop-Off (PUDO) areas	With the increase in demand for private hire cars today, existing Pick-Up Drop-Off (PUDO) at various developments could have difficulty coping with increased flows, resulting in traffic spill overs. There are limited studies on the factors that influence vehicular flows into PUDOs. We are interested to quantify current and project future PUDO volumes to inform the future-proofing of PUDO design. The objective of this project is to design a user-friendly model to predict vehicle flows and spill-overs for several typologies of PUDOs.	Basic modelling skills, R/Python (Fieldwork may be necessary to collect additional data)
2	URA	Developing a Video Analytics algorithm to detect conflict between active mobility users	As part of our move towards car-lite, active mobility modes are highly encouraged. Pedestrians, cyclists and PMD users often meet along shared paths and traffic junctions. Thus, there is a need to ensure that we provide comfortable and safe walkways for all users. An understanding of user interaction and conflict will be important to translate into useful applications. Using footage collected along a footpath, the study would tag all interactions between active mobility users observed. It will then measure the angular change in direction and change in speed to derive some threshold values that determine conflicts. Applying machine learning, an algorithm should be designed to automatically pick out conflict cases in video and categorise these cases.	Machine learning/Neural network, Python
3	*Scape	Smart Toilet Management	*SCAPE is a 5 storey building with a total of is 6 individual toilets for those on wheel-chair, 6 common toilets for the male and 6 common toilets for the female with varying number of cubicles between 4 – 10 cubicles inside. The toilets’ lights and fans are switched on, in most of the toilet, during the building operation hours from 8am to 12pm for levels 2-5. Except for 1 individual toilet for those on wheel-chair, 1 common toilet for the male and 1 common toilet for the female located on the 2nd level which is switched on 24 hours a day and 7 days a week. At times, during events, peak seasons and peak hours, there were several feedback that the toilet are not cleaned and supplies such as the toilet paper are not stock up. Estates team, sometimes, needs to visit to verify the feedback and or get the cleaners to attend to the feedback. Estates is exploring whether sensors can be deployed to: turn on the lights and fans upon detecting movement into the toilet; turn off the lights and fans if no movement is detected over a certain period of time; detect the presence of a certain constant level of CO2 over a certain period of time and to notify the cleaners; notify the cleaners to clean up the toilet after a certain number of visits recorded for that particular toilet; notify the cleaners that toilet roll in certain cubicle is running low; & notify the cleaners that the bin is full and has to be emptied. Objectives of the project include: Electricity can be saved by switching off the lights and fans in those toilets that are not in use. Thus a saving in real costs in utility and also in procurement from a longer life span of the lightings and fans ; To help contribute towards a Greener Environment; The cleaners can be redeployed for other cleaning matters if the toilets does not warrant to be clean for the time being; Efficiency of the cleaners will be improved based on the notification from the sensors deployed and Estates team can focus on other pressing areas; A cleaner environment for the users; Portraying a better image for the building; & A step towards a smart facility management system	IoT
4	SCDF	Fire Safety in Singapore	The Fire Safety Act requires buildings in Singapore to be equipped with Fire Alarm Systems, which are connected to smoke or heat detectors that will trigger the alarm in the presence of said by-products of fire. However, often times the detectors do not get triggered until the fire has been well developed. This could be due to the slow rise of temperature, of if the smoke is emitting away from the detector. By the time the fire alarm rings, half the place might have been burnt down. The idea is to think of an innovative yet cost efficient way to detect the presence of a fire when it is still in its early stage. Currently there are solutions such as thermos cameras that can detect small fires but they cost in the thousands. Quite unlikely that companies would want to purchase these cameras.	IoT