University research in Toronto, Canada.
Nationwide, United States
Using Microsimulation to Evaluate the Impact of Automated Vehicles on Safety Performance of Signalized Intersections
Summary Information
This study used a micro-simulation model to generate simulated traffic conflicts as indicators of potential crashes at signalized intersections in Toronto, Canada in the presence of automated vehicles (AV) at various penetration levels.
METHODOLOGY
Seventy-eight (78) 4-legged signalized intersections from the city of Toronto were first modelled in the SYNCHRO software. Ten random runs for 3600 seconds (representing peak hour) were simulated for each intersection. The steps mentioned above were applied to the three penetration levels of automated vehicles investigated (0%AV, 50%AV and 100%AV).
All automated vehicles were assumed to have the same level of automation (Level 3 per NHTSA definitions).
FINDINGS
The results indicate that intersection safety may improve in the presence of AVs. However, the safety effects of treatments (intersection design and signal timing treatments for turning movements) may be reduced compared to the effects with no AVs. The implication is that the imminent introduction of AVs should be considered in developing priorities for future intersection improvements. The data in the table below were excerpted from the source report.
Crashes per intersection-year at various AV market penetration levels as estimated from crash-conflict models.
Percentage market penetration |
Total crashes using |
Injury crashes using |
Angle crashes using |
Rear-end crashes using |
---|---|---|---|---|
0% |
33.94 |
7.42 |
5.36 |
10.60 |
50% |
25.66 |
5.81 |
5.30 |
7.64 |
100% |
24.74 |
5.62 |
5.37 |
7.23 |