This document summarizes research into developing an intelligent system to detect cyclists at roundabout junctions and generate alerts. Key points: - Researchers at TSSG tested using Bluetooth beacons and GPS to detect cyclist trajectories and send alerts to different zones of the junction. - Bluetooth detection worked well at speeds of 15-20 km/h when beacons were at a height of 1.2m, accurately tracking cyclist movement. - GPS was deemed a more stable and lower cost solution than Bluetooth. Internet connectivity is important for the GPS-based system. - The system was able to generate different alerts to different junction zones depending on detected cyclist movement, but more real-world testing

1. Intelligent and Safer Junctions,
Supporting Safer Cycling
3rd – 4th November, 2016 @ Bolonga Italy
John McLaughlin jmclaughlin@tssg.org, Dr. Steven Davy sdavy@tssg.org,
Frances Cleary fcleary@tssg.org, David Ryan dryan@tssg.org
Programmable Automomous Systems Unit

2. Introduction,
• About TSSG: TSSG is a telecommunications research
institute attached to Waterford Institute of Technology,
established in 1996, in South East Ireland.
• Aim: The aim of the project was to research a sensor based
system to detect the presence of cyclists at roundabout
junctions and devise a system to generate alerts specific to
the affected area on the junction.
www.tssg.org 2
Intelligent and Safer Junctions, Supporting Safer Cycling

3. Key Outputs
• To assess if Bluetooth Beacon Technology is suitable for
cyclist detection at roundabout junctions.
• To develop a low cost, light weight system for creating
alerts at the separate junctions on a roundabout based
on the cyclist’s trajectory.
www.tssg.org 3

4. Safer Cycling Use Case
Cyclist Path
Car Path
Collision Points
Detection and Early Warning
4

5. System Architecture
15-Nov-16

6. Cyclist Location Sensing Beacon Approach
• Tags Strategically Positioned
• Signal Received by smart device carried by cyclist
• The position of cyclist is computed by trilateration.
• The position of the cyclist to the Bluetooth tag is determined
• When the position of the cyclist is determined in relation to at
least 3 tags the position of the cyclist is known by the system

7. GPS Location Sensing Approach
• This method exploits GPS technology on device held by the
cyclist.
• A manifest of junctions is sent to the mobile device from the
Listening Service.
• The app determines, using mathematics, the current GPS
position and the manifest, the relevant junction and once
within a nominal distance the application begins to stream the
current GPS position to the Listening Service
www.tssg.org 7

8. Junction Zones
• Virtual Zones added to
Junction using a purposely
built online map tool
• These zones are used to
define the transition of the
cyclist on the junction within
the system.
• They form the basis of the
policy and alerting system

9. The policies are a set of rules that,
based on a transition
between zones define what
alerts are sent to the junction
Policies & Alerts
Rule : [Zone 2 -> Zone 3]
ALERTS : [ Zone 5 -> Zone 7 ]
Rule : [Zone 3 -> Zone 7]
ALERTS : [ Zone 8 -> Zone 10 ]
CANCEL : [Zone 2 -> Zone 3]
Rule : [Zone 7 -> Zone 10]
ALERTS : [ Zone 2 -> Zone 3 ]
CANCEL : [Zone 5 -> Zone 7]

10. Findings
• At speeds of 15km and 20km the cyclist was detected.
• The height of the Bluetooth Beacons were crucial to the detection
with 1.2m being the optimum.
• Sufficient rate of detection was achieved by this method to
accurately detect the cyclist’s trajectory through the junction.
• The detected trajectory determined by trilateration was similar to the
trajectory determined by the Garmin 500 unit that was used as a
comparison.
Bluetooth Beacon Technology Assessment

11. Findings
• The key component to the system was the zone
transition and policy system.
• GPS Based Location sensing was deemed a more
stable and lower cost solution to the Bluetooth Beacon
Technology
• GPS signal and adequate internet connectivity is key to
the success of the system.
Development of low cost cyclist detection system

12. Conclusion
• For 100% cyclist detection it requires all cyclists to use the
system i.e. use one of the cyclist detection mechanisms.
• Extra trials required to test the system in a real world scenario
to determine if the system would be suitable in a complex
urban environment.
• From a proof of concept perspective the system provides a
low cost, light weight solution to making roundabout junctions
more intelligent and potentially safer for cyclists

13. DELIVERING SCIENCE TO SHAPE
YOUR FUTURE.
Any questions?