1 of 15

2 of 15

Perception-Based

Adaptive Traffic Management

Jamie Mackey | Olsson

ITS Arizona | October 9, 2025

3 of 15

Can your traffic signal see?

No detection

“Blind” - No real-time inputs
Common operation

Traditional detection

“Raise hand” - On/off at static locations
Emulate traffic conditions
Considered great operations

Real-time data stream from sensor

Speed, position, direction, classification, ETA
Connected vehicle will do this, but when?

4 of 15

DETECTION is like standing behind a wall and only seeing traffic that passes in front of window

PERCEPTION is like stepping around the wall to see all traffic unobstructed

Detection vs Perception

5 of 15

How it Works

Diagram from Colorado Springs SMART Grant report.

Redundant Sensors: �Radar, video, LiDAR, weather, multimodal, connected vehicle

Data Fusion into a Digital Twin: Combining inputs to create a live model of the intersection

Controller Firmware Evolution: From binary inputs to multi-dimensional perception

6 of 15

Detection – What the Controller Sees

Traditional Controller Inputs

Cloud-based

Data Processors

Cabinet-based

Data Processors

7 of 15

Detection – What the Controller Sees

Traditional Controller Inputs

Traffic Signal

Controller

Cloud-based

Data Processors

Cabinet-based

Data Processors

8 of 15

Detection – What the Controller Sees

Traditional Controller Inputs

Traffic Signal

Controller

Cloud-based

Data Processors

Cabinet-based

Data Processors

9 of 15

Detection – What the Controller Sees

Traditional Controller Inputs

Traffic Signal

Controller

Cloud-based

Data Processors

Cabinet-based

Data Processors

10 of 15

Detection – What the Controller Sees

Traditional Controller Inputs

Traffic Signal

Controller

Cloud-based

Data Processors

Cabinet-based

Data Processors

11 of 15

Detection – What the Controller Sees

Detailed Controller Inputs

Traffic Signal

Controller

Cloud-based

Data Processors

Cabinet-based

Data Processors

12 of 15

Safest time to end greens: Dilemma zone protection based on vehicle size and speed *

Rest-in-red speed compliance *

Conflict-based red extensions

Vulnerable road users
Red light runners

Platoon progression for smoother flow *

Green extension to reduce split failures

Balance delay or queuing during congestion

Safety

Efficiency

Applications of Perception-Based Operations

Diagram from Colorado Springs SMART Grant report.

* Strategy deployed

13 of 15

The City of Colorado Springs is committed to making the work available through �open-source platforms and crowdsourcing.
Pooled Fund Study TPF-5(483)
$1.7 million SMART Grant
NTCIP compliance coming soon

The Project: Colorado Springs SMART Grant

Diagram from Colorado Springs SMART Grant report.

14 of 15

Questions?�

Jamie Mackey

jmackey@olsson.com

1 of 15

2 of 15

3 of 15

4 of 15

5 of 15

6 of 15

7 of 15

8 of 15

9 of 15

10 of 15

11 of 15

12 of 15

13 of 15

14 of 15

15 of 15