Case Study: Quickly improving traffic progression in the Town of Milton

When citizens in the Town of Milton reported complaints about poor traffic progression at two busy intersections, the town’s traffic engineering team sought a cost-effective way to analyze and resolve the situation. Using plug-and-play traffic solutions, it took only four weeks to analyze, diagnose, and resolve the issues, resulting in:

  • A 20 to 33% reduction in split failures, which addressed the top cause of citizen complaints
  • A 20% reduction in travel time, attributable to less-interrupted flow along the corridor
  • Demonstrably less vehicle queuing
Milton case study video

The Town of Milton, a community west of Toronto (and one of the city’s largest commuter communities), was fielding citizen complaints about the performance of two important intersections on Main Street.

Aware of the shortcomings of manual counts, the town’s traffic engineering team needed a solution that could generate reliable data about traffic signal performance. Using Miovision TrafficLink, the team was not only able to collect high-quality (and auditable) data, but they were also able to implement corrective action and measure the impact of those changes to ensure there were noticeable improvements for citizens.
Using signal performance measures to identify problems. Milton installed Miovision SmartLink, SmartSense, and SmartView 360 at the two key intersections and using TrafficLink Performance Measures, generated insight reports that revealed:

  • Unexpected variations in cycle length at one of the intersections
  • Extremely high occupancy ratios
  • A high number of split failures

Visual examination using Miovision SmartView 360 cameras also showed that the major approaches were all experiencing high volumes of traffic, which caused traffic to spill back onto upstream intersections.

Uncovering the root cause of traffic congestion

One of the issues identified using TrafficLink was the cause of the irregular change in cycle lengths. This irregularity was the result of a controller going into a transition mode. Investigation revealed that this variable performance was caused by an error in the controller’s documentation, which was reflected in the signal timing plan.

Traffic engineers updated the signal timing plan accordingly and the inconsistent cycle lengths were resolved. This simple correction improved signal performance by dropping split failures by 8.5%).

graph showing inconsistent cycle lengths

The graph above shows the inconsistency cycle lengths due to the controller’s documentation error. The graph below shows the more consistent cycle lengths after correcting the signal timing plan.

Graph showing consistent cycle lengths after correcting the signal timing plan

Investigating split failures and progression challenges

A split failure occurs when a phase cannot serve all its demand within one cycle; that is, if it takes a vehicle two or more cycles to execute its movement at an intersection, then a split failure has occurred. Split failures cause frustration among drivers and were the major reason for complaints received by the Town of Milton.

The Purdue Split Failure is an industry-standard metric that charts the frequency of split failure occurrences. TrafficLink showed split failures during the midday period on the southbound turning movement. However, other phases were undersaturated at this time, which suggested a potential corrective action: reallocating some green time from the main movement to the southbound turning movement.

Another issue revealed by TrafficLink was progression challenges; closer inspection using the video footage revealed green starvation on eastbound turning movements. Additionally, during the busy evening peak, the downstream intersection is green while there is no demand to serve, wasting green time.

A time-space diagram of the timing allocations showed poor coordination between the intersections, creating a long queue of vehicles and spillback.

Finding a solution and seeing the results

Based on the data presented in TrafficLink, the traffic engineering team implemented a very small offset (5 to 6 seconds) between the two intersections. The team then used Automated Traffic Signal Performance Measures (ATSPMs) to evaluate the impact. A before-and-after analysis showed that this small change had a very large and positive impact on the performance of the intersections:

  • Travel time along the corridor was reduced in both directions
  • Split failures were cut substantially

Note: These results were achieved without changing the green time allocation — the slight offset introduced by the engineers had the effect of creating enough room for vehicles to use the green time at maximum capacity, the end result of which was allowing more vehicles to transit, thereby reducing the number of split failures.



The two Purdue Split Failure charts above show the split failures during the evening hours. The first chart shows the “before” state, while the second chart shows a 20% reduction in split failures after introducing a six-second offset between the two intersections.