What is a near-miss?

A near miss is a statistically significant speed and g-force event. It shows where drivers are losing friction with the road.

A near miss is a statistically significant speed and g-force event that falls under one of three categories of causation: braking, steering, or a combination of both. This algorithm is based on substantial road safety literature, and a series of real-world experiments.

A near-miss acts as a leading indicator, identifying where potentially high-risk intersections or roads exist across a transport network. 

It typically implies that the vehicles involved were in a potentially dangerous situation where an accident could have occurred, but due to timely actions by the drivers or other factors, a collision was prevented. Despite the absence of a collision, a near miss is considered a serious incident that highlights the importance of safe driving practices and situational awareness on the road.

There is a difference to the typical definition of a near miss and how near miss is described in Compass data. Compass refers to near misses as:

An occurrence involving a rapid evasive maneuvre (from braking or steering or a combination of both)

This means that the near miss values shown in the data outputs indicate that above or below a certain threshold the ‘near miss’ algorithm is triggered as the vehicle is about or has lost friction with the pavement during a braking, accelerating or swerving event.

The ‘near miss’ algorithm uses a range of different acceleration (g force) and speed thresholds depending on the vehicle type i.e. whether it is a HCV, LCV, car, or bus. These thresholds are shown in the next slide

The Role of g-forces

A “G-Force” is a measure of acceleration. 1 G is the acceleration due to the force of gravity. It is measured in m/sec2. The acceleration of braking and swerving is the equivalent force to gravity. G-forces provide a measure of the intensity of a near-miss and the direction or data (i.e., left vs right or acceleration vs deceleration)

In the Compass data, there are 3 types of g-forces showed in the Compass outputs:

  • Braking: -x
  • Accelerating: + x
  • Swerving: – y (left swerving) and + y (right swerving)
The main purpose of these incidents (near misses) are not to identify actual crashes but to identify locations where the road network fails, either due to design issues, or traffic operational issues or driver behaviour