Technical Name Applying Machine Learning to User Mobility Type Identification for 5th Generation Mobile Networks
Project Operator National Chiao Tung University
Project Host 陳志成
Due to the fast development of 5G networks, it is critical to identify users service types to allocate resources intelligently. Our technology focuses on users mobility type identification by extracting practical features from users cellular information. We proposed a system architecturehave collected 700-hour data with 150 GB. By using our dataother datasets in the world, we show that our technology can achieve 95 accuracy,reduce 16 energy consumption compared to traditional methods.
Technical Film
Scientific Breakthrough
Mobility type identification using smartphone sensors has some limitations:

1. Environment: GPS may be blocked by obstacles,magnetometer, barometerlight sensor can be affected by environments. 
Solution: Our technology applies cellular data instead of using those sensors.

2. Energy consumption: Computing resources limit the long-term sensing by smart devices. 
Solution: Our technology only needs cellular information with event-based sampling to reduce energy consumption.

3. Privacy issue: Smart device sensors may leak users locationphysical activity information. 
Solution: We only collect cellular informationit has coarser granularity of location information.

This technology can achieve 95 accuracyreduce 16 energy consumption compared to traditional methods.
Industrial Applicability
1. Network providers: Our technology is designed for 5G networks to identify users mobility types, so that network providers can deploy specific mechanisms for users on high-speed transportation,offload network traffic from car-to-outside-BS to car-to-inside-BS.

2. People: Our technology can be deployed as a mobile app on smartphones for smart navigation, carbon footprint, elderly tracking,calories calculator.

3. Vehicles: This technology can be used for usage-based pricingdriving behavior analysis, allowing service providersinsurance companies to innovate.

4. Cities: This technology can be used to understand the mobilitytraffic patterns in cities for applications such as congestion control, traffic planning,travel time prediction.
Matching Needs
1. Mobile device applications development vendors
2. Mobile communication analysis
3. Mobile network operator
Keyword Transportation Type Identification Mobility Type Identification Machine Learning Deep Learning Classification Mobile Network 5G Mobile Network Cellular Information Mobile Crowd Sensing Smart City
Download 08761519.pdf
TCSE2020_Demo_具社交功能的移動群眾量測平台SensingGO實作展示 (1).doc
  • Contact