Crash With

Passenger Car

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Making Right Turn

SV Contact Area

Rear

SV Precrash Speed (MPH)

2

CP Pre-Crash Movement

Making Right Turn

CP Contact Area

Front Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Investigating Agency

East District Arlington Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May Mobility vehicle (Ego) is stopped at a red light, then the Ego begins to accelerate and turn right when the traffic light changes to green. As the Ego accelerates, a pedestrian begins moving and enters the crosswalk from the other side of the intersection. This activates the rule-based crosswalk logic. Ego determines that it is able to stop before the crosswalk without exceeding our max deceleration threshold, which triggers an in-policy stopping reaction for the crosswalk. The Ego brakes to yield for the pedestrian. The vehicle trailing the Ego accelerates, making contact with the rear-end of the Ego. In the 2.5 seconds between the start of the Ego braking and the collision, the rear vehicle accelerated from 0.2 to 6.1 m/s (0.4 to 13.6 mph) without slowing until after the collision occurred. In response to this incident, immediate mitigations were made, in that the vehicle operators (AVOs) operate all intersections in manual mode. This remained in place until January 12, when we received a solution which instituted a reduction of the 3.5 m/s2 parameter, in order to prevent the Ego from yielding to pedestrians in crosswalks that are still far from the Ego’s lane if such yielding would induce braking.

Crash With

Passenger Car

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Making Right Turn

SV Contact Area

Rear

SV Precrash Speed (MPH)

4

CP Pre-Crash Movement

Proceeding Straight

CP Contact Area

Front Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May Mobility vehicle (Ego) approaches a turn with a trailing vehicle following closely behind. The May planner initially selects a trajectory it believes will safely complete the turn. However, as the Ego begins the turn, it re-evaluates the situation and detects a significant risk of striking the curb. At that point, the planner shifts focus from avoiding the curb entirely to minimizing the severity of a potential curb strike. To reduce speed before the expected contact, the Ego brakes at €“3.5 m/s²). Because the trailing vehicle is following so closely, this deceleration makes a rear-end collision likely. The Ego predicts a potential rear-end collision and elects to reduce brake force to lessen the impact severity. Moments later, the trailing vehicle strikes the rear driver-side bumper of the Ego and then drives away. Future improvements are underway to give the planner a policy option to steer away from curbs. This new capability will help the vehicle avoid situations where braking becomes the only remaining way to manage close curb proximity.

Crash With

SUV

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Proceeding Straight

SV Contact Area

Right

SV Precrash Speed (MPH)

5

CP Pre-Crash Movement

Changing Lanes

CP Contact Area

Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Street

Weather

Cloudy

Investigating Agency

Atlanta Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May Mobility vehicle (Ego) was operating in autonomy while waiting in slow traffic in the leftmost of two parallel lanes. Ahead, both lanes were obstructed by stopped vehicles, prompting traffic to move to the center turn lane to proceed. The Autonomous Vehicle Operator (AVO) intervened to move closer to the lead vehicle as another vehicle in the right lane attempted to merge left, preventing it. After the merging car stopped, the AVO pulled closer to the lead vehicle, removing any space to merge. As the lead vehicle began to move, both the AVO and the vehicle to the right immediately accelerated, while the merging vehicle moved left into the side of the May AV. Had the Ego remained in Autonomous mode, it would have maintained a safe distance allowing the merging vehicle to proceed safely. No software defects led to this event. This behavior goes against the training of May AVO€™s. Upon review of this AVO’s actions it was decided to remove this person from their role as an AVO.

Crash With

Other Fixed Object

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Proceeding Straight

SV Contact Area

Right, Bottom

SV Precrash Speed (MPH)

5

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Parking Lot

Weather

Cloudy, Rain

Investigating Agency

Atlanta Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

During normal autonomous operation, the May autonomous shuttle (Ego) was traveling through a narrow section of roadway in a parking lot bordered by parked vehicles on the left and a retaining wall on the right on experimental software version v29.0.20. As the Ego proceeded, the system initially recognized the static retaining wall as a large, non-rectangular object. As the Ego approaches the wall, the resulting density of lidar returns on the object (wall) exceeds the detectors maximum limitation. The newer lidar points, toward the forward edge of the elongated object, exceeding the limit were filtered out, causing the track to disappear when the ego approached 5m from the track (i.e. the wall). The situation was exacerbated by the fact that the road edge was incorrectly marked in that section. The road edge was drawn through the retaining wall and thus the vehicle did not have any prior knowledge about the edge of the road. As a result, the Ego drove too close to the wall, scraping it at a speed of 2 meters per second while driving straight (i.e. in lane-nominal policy) with no nudge or veer corrections. There was no damage to the wall. The limitation leading to the lost track affects Software versions v26.0 (when the Open World Tracklet v2 feature was introduced) through v31.0. This is the only known instance of this failure mode, and in v32.0 an optimization was made to remove the LIDAR threshold entirely and mitigate the vulnerability. Furthermore, the incorrect road edge annotation was corrected shortly after this incident and the map updated for the deployed vehicles.

Crash With

Passenger Car

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Stopped

SV Contact Area

Front

SV Precrash Speed (MPH)

1

CP Pre-Crash Movement

Backing

CP Contact Area

Rear Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Investigating Agency

Atlanta Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May autonomous shuttle (Ego) stopped in Autonomous mode safely behind a lead vehicle at a traffic light, which then, the lead vehicle, reversed into the stationary Ego. The Autonomous vehicle operator took over control of the vehicle and attempted to reverse, but was unable to get out of the way of the reversing vehicle. The planner did roll out the agent with a reversing policy and predicted the collision, but the system currently is not able to honk or reverse or do anything else that could have avoided this. The primary factor was fault by the lead vehicle. There was no fault in our software that led to this event.

Crash With

Other Fixed Object

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Parking Maneuver

SV Contact Area

Front Left

SV Precrash Speed (MPH)

25

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Street

Weather

Clear

Investigating Agency

None

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

A May autonomous shuttle (Ego) is approaching a passenger stop location in autonomy. The vehicle operator (AVO) presses on the accelerator, taking control of the Ego, then accelerates into the curb at the stop. The Ego made contact with the curb 3 seconds after switching into manual, causing severe damage to the front-right tire. Resimulation of this event shows that there were no autonomy errors that led to this event, and had the Ego stayed in autonomy, this event would not have happened. This AVO had just completed our training course with only a few hours of service time before this incident. Upon further review of the AVO performance, on their own, it was decided to remove her from the AVO role.

Crash With

Passenger Car

Highest Injury Severity Alleged

Minor W/O Hospitalization

SV Pre-Crash Movement

Proceeding Straight

SV Contact Area

Front Right

SV Precrash Speed (MPH)

13

CP Pre-Crash Movement

Making Right Turn

CP Contact Area

Rear Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Investigating Agency

Gwinnett County Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May vehicle was travelling in an autonomous vehicle (AV) only lane approaching an intersection. A vehicle travelling to the left of the May shuttle (ego) in the non-AV lane performed a right hand turn across the May vehicleƒÂ¢Ã‚€Â™s path. Given the full stopping power and the limited space with which the ego had to stop, the ego was physically unable to stop in time before impact. The primary factor that led to this collision was human error on the part of the other vehicle that was required to yield to through traffic in the AV lane but did not. Right turns across a straight lane are uncommon, and drivers are typically not used to them, however there are several signs that instruct other traffic to yield to the vehicles in the AV lane. There were no autonomy faults or errors which led to this incident. Update: There was no indication of injury or request for medical attention at the scene. The police report states "Driver 1 complained of neck pain." For this reason, we are updating the highest severity injury alleged field to reflect the “minor w/o hospitalization” injury level. The remaining content of this report is unchanged from the initial report submitted on September [XXX], 2025, Report ID #[XXX].

Crash With

Pickup Truck

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Proceeding Straight

SV Contact Area

Rear Right

SV Precrash Speed (MPH)

13

CP Pre-Crash Movement

Proceeding Straight

CP Contact Area

Front Right

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Investigating Agency

Eden Prairie Police Dept

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

A May autonomous shuttle (Ego) is passing through an intersection. A vehicle sensor detects a false positive as a pedestrian moving toward the May vehicle€™s path in a region closely behind another vehicle. This causes the May vehicle to stop to allow it to safely pass. A following agent (still accelerating at the time of the impact) collides with the May vehicle one second after we come to a stop. Re-simulation on newer software shows that routine software improvements resolved this issue and the commercial May software has been updated to this release.

Crash With

Passenger Car

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Proceeding Straight

SV Contact Area

Front

SV Precrash Speed (MPH)

20

CP Pre-Crash Movement

Making Right Turn

CP Contact Area

Front Left

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Intersection

Weather

Clear

Investigating Agency

Gwinnett County Police Department

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

The May autonomous vehicle was travelling in an autonomous vehicle (AV) only lane, on [XXX], approaching the intersection at [XXX] A vehicle travelling to the left of the May shuttle in the non-AV lane attempted to perform a right hand turn across the May vehicles path and collided with the May AV. The primary factor that led to this collision was human error on the part of the other vehicle that was required to yield to through traffic in the AV lane but did not. Right turns across a straight lane are uncommon, and drivers are typically not used to them, however there is a metal street sign that instructs other traffic to yield to the vehicles in the AV lane. There were no autonomy faults or errors which led to this incident. Additionally, there is a second traffic light at this intersection that warns public drivers to yield to the AV shuttle. This light was not operating as it should during this incident. We are working with our partner to maintain the operation of this light and have implemented a mitigation policy of daily inspections of this light and implementing an AVO in service if this light is not functioning.

Crash With

Other Fixed Object

Highest Injury Severity Alleged

No Injured Reported

SV Pre-Crash Movement

Making Right Turn

SV Contact Area

Right

SV Precrash Speed (MPH)

18

Within ODD?

Yes

Automation System Engaged?

ADS

Roadway Type

Street

Weather

Clear

Investigating Agency

None

Make

TOYOTA

Model

Sienna

SV is the subject vehicle — the one whose operator filed the report. CP is the crash partner, the other party. Contact area is where each was struck. NHTSA requires a report for any crash involving the driving system regardless of fault.

While traveling north on [XXX], the Autonomous Vehicle Operator (AVO) engages autonomy 3.0 m before the planned right turn and 7.7 m before the break in the curb while traveling at 8.0 m/s. The May Autonomous Vehicle (ego) was planning to make a tight right turn into a parking lot, which the AVO was unaware of when engaging autonomy. The ego jerked to the right upon entering autonomy, hitting a curb and damaging both passenger-side front and rear tires and the right rear tire rim. The AV attempted the right turn despite not having time to slow sufficiently and mounted the curb. The AVO intervened by accelerating and steering back to the left upon striking the curb, manually maneuvering back to the street while striking a second curb on the opposite side of the parking lot driveway. Resimulations without the manual intervention show the ego slowing appropriately and making the right turn into the parking lot. The AVO should have been attentive to the planned route before engaging autonomy. As a result of this incident, planned improvements to the AVOs instructive dash and audio alerts for upcoming turns have begun to be implemented.