News listWaymo autonomous vehicles repeatedly drive into flooded areas, Robotaxi operations suspended in four cities
動區 BlockTempo2026-05-22 03:11:27

Waymo autonomous vehicles repeatedly drive into flooded areas, Robotaxi operations suspended in four cities

ORIGINALWaymo 自駕車反覆衝進淹水路段,Robotaxi 擴大停運至四城
AI Impact AnalysisGrok analyzing...
📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1891 words
Waymo has initiated a recall of its 3,791 robotaxis and suspended services in four cities, including Atlanta, due to the vehicle software's inability to identify flooded roads. Following an incident in San Antonio in April where a vehicle drove into a creek, a similar issue occurred in Atlanta on May 21. (Previous coverage: Waymo blocked an ambulance rushing to a shooting scene; emergency responders: the situation is getting worse) (Background: Tesla Robotaxi begins trial operations in Texas; autonomous taxis to challenge the status quo of Waymo and Uber) In Silicon Valley labs, driving technology deals with complex decision-making, such as identifying pedestrians and predicting behavior at intersections within milliseconds. However, after a heavy rainstorm in Texas, Waymo's robotaxi stumbled over a much simpler problem. The incident began on April 20. A Waymo robotaxi drove into a flooded section in San Antonio, Texas, and was eventually swept into a creek. This accident triggered the attention of the NHTSA (National Highway Traffic Safety Administration) and forced Waymo to confront structural issues within its fleet software. On May 12, Waymo initiated a recall of its 3,791 robotaxis. This was not a traditional recall requiring owners to drive vehicles back to a service center, but rather an OTA (over-the-air) update, where vehicle software is pushed remotely via the internet without the need for a shop visit. Waymo admitted that this update was merely a transitional measure: restricting vehicle operations in "high-risk flood areas during specific time periods," while the "final solution" has yet to be completed. Before the dust could settle, another incident occurred in Atlanta on May 21. An empty Waymo robotaxi drove into a flooded street, became stuck for nearly an hour, and required a tow truck to be recovered. The rainfall that day was heavy enough to cause street flooding, but the National Weather Service had not yet managed to issue a flood warning. This means that the "high-risk time restrictions" set by Waymo's previous OTA update were completely ineffective during this rapidly developing regional rainstorm. Following the incident, Waymo announced the suspension of services in four cities: Atlanta, San Antonio, Dallas, and Houston. To understand why Waymo repeatedly fails on the same issue, one must return to the technical essence of autonomous vehicles. The perception system of a robotaxi relies on training with massive amounts of labeled data: where the drivable surface is, where obstacles are, and which lines mark lane boundaries. In the distribution of training data, city streets are "dry, normal roads" 99% of the time. Flooded sections, in the language of machine learning, are edge cases: rare scenarios not sufficiently covered by training data. The problem is that when a robotaxi's sensors (cameras + LiDAR) scan a pool of water, the reflective properties of the water are very similar to wet asphalt. Without being explicitly taught that "this scenario represents danger," the model is not "making a judgment error"—it simply wasn't designed to brake in this scenario. Waymo's OTA update attempted to circumvent this issue using "geofencing + time restrictions": prohibiting or limiting vehicle operations in known high-risk flood areas during specific rainy periods. This logic itself is sound, but it relies on a premise: the early warning system must be faster than reality. The Atlanta case on May 21 proved that the speed of a rainstorm can outpace NWS alerts and, even more so, the trigger logic of geofencing. Waymo's commercial operations cover 11 U.S. cities, including San Francisco, Los Angeles, Phoenix, Austin, and Miami. The suspension in four cities means that more than one-third of its operational footprint has entered a state of pause. This is not just a regional issue, but a systemic risk brought about by the entire fleet sharing the same software logic. The flooding incidents are not the only regulatory pressure Waymo currently faces. Beyond the flooding recall, the NHTSA and the NTSB (National Transportation Safety Board) are jointly investigating another persistent issue: Waymo robotaxis repeatedly and illegally passing stopped school buses. State laws across the U.S. require vehicles to stop and wait beside school buses with extended stop arms, but Waymo vehicles have violated this rule multiple times in Austin. Even after a fix was pushed, the problem persisted, leading the NHTSA to issue a second round of information requests on May 15. During the same period, on January 23 of this year, a Waymo robotaxi struck a child at a low speed of approximately 10 km/h in Santa Monica, Los Angeles, causing minor injuries. The incident occurred near an elementary school. In a statement, the NHTSA indicated that it "will take action if necessary" regarding the flooding incidents. The
Data Status✓ Full text extractedRead Original (動區 BlockTempo)
🔍Historical Similar Events· Keyword + Asset Matching0 items
No similar events found (requires more data samples or embedding search; currently MVP keyword matching)
Raw Information
ID:56adc50f27
Source:動區 BlockTempo
Published:2026-05-22 03:11:27
Category:zh_news · Export Category zh
Symbols:Unspecified
Community Votes:+0 /0 · ⭐ 0 Important · 💬 0 Comments