Zoox Recalls Software After Autonomous Vehicle Drives Into Heavy Smoke

Zoox, the autonomous vehicle subsidiary of Amazon, has issued a recall for software in 105 of its robotaxis. The company proactively alerted the National Highway Traffic Safety Administration (NHTSA) on July 8 regarding a critical issue where its vehicles failed to detect heavy smoke, leading to them entering active emergency scenes. This recall underscores the complex safety challenges inherent in the rapid development and deployment of autonomous driving technology.

The incident that triggered the recall occurred on June 20 in Las Vegas. An unoccupied Zoox robotaxi encountered dense smoke, which had obscured an emergency fire scene that was not yet secured with traffic cones. The vehicle proceeded into the scene before abruptly braking and attempting to steer away, ultimately coming to a standstill. A Zoox teleguidance employee then remotely instructed the vehicle to reverse. Following this, first responders placed traffic cones to delineate the hazardous area.

Zoox has stated that its internal investigation identified this as the sole occurrence of this nature and reported no injuries resulting from the incident. This recall, though voluntary, comes at a pivotal moment for the autonomous vehicle industry, which is under increasing scrutiny from regulators and the public.

Amazon acquired Zoox for $1.3 billion in 2020, aiming to accelerate its ambitions in the autonomous mobility space. Zoox’s vehicles are designed for urban environments, featuring a unique bidirectional design without traditional steering wheels or pedals, and inward-facing carriage-style seats, creating a communal, shuttle-like experience for passengers. Currently, Zoox offers complimentary rides in select areas of Las Vegas and San Francisco, and is piloting limited services in Miami and Austin, Texas, with extensive testing underway in an additional six U.S. cities. This expansion strategy, while aggressive, highlights the company’s commitment to scaling its operations amidst fierce competition.

The timing of Zoox’s recall is particularly noteworthy, as it follows a recent directive from NHTSA Administrator Jonathan Morrison. Last week, Morrison issued a formal letter to autonomous vehicle developers, urging them to prioritize solutions that ensure their vehicles yield to first responders. The NHTSA has observed a “clear pattern” of driverless vehicles interfering with emergency services, citing instances where AVs have entered active scenes, obstructed ambulances or fire trucks, or failed to recognize critical visual cues like flashing lights, flares, smoke, and cones. The agency has mandated that developers present their remediation plans by the end of July.

This latest recall adds to Zoox’s history of addressing software-related safety concerns. In the past year, the company has issued recalls for issues related to lane crossings, the prediction of other vehicles’ movements, and pedestrian detection. These recurring challenges point to the intricate nature of developing robust perception and prediction algorithms capable of handling the myriad unpredictable scenarios encountered on public roads.

Zoox is in a race to establish a significant market presence, facing formidable competition from established players like Alphabet’s Waymo. Waymo currently commands the dominant position in the U.S. robotaxi market, operating a fleet of approximately 4,000 autonomous vehicles. Just last month, Waymo itself recalled nearly 3,900 robotaxis due to incidents where vehicles entered closed freeway construction zones, increasing crash risks. The competitive landscape is characterized by a relentless drive for technological advancement and operational efficiency, with safety and regulatory compliance being paramount to sustained success.

The industry’s focus is increasingly shifting towards demonstrating not just technological prowess, but also unwavering reliability and public trust. As autonomous vehicle technology matures, the ability of companies like Zoox to effectively address complex safety issues, particularly those involving interactions with emergency services, will be a critical determinant of their long-term viability and widespread adoption.