After years of promising investors that millions of Tesla Robotaxis would soon flood the streets, Elon Musk launched a limited driverless car service in Austin, Texas. The rollout faced significant challenges from the start.
The June 22nd debut was met with a barrage of videos from pro-Tesla influencers, who appeared to celebrate the service and showcased their rides. Musk heralded it as a milestone, and Tesla’s stock shot up nearly 10% the next day.
However, it soon became evident that some of the influencer footage painted a troubling picture of an autonomous vehicle that either broke traffic laws or struggled with basic functions. By Tuesday, the National Highway Traffic Safety Administration (NHTSA) had launched an investigation into these incidents and sought Tesla’s input.
If, as Musk boasted on X, this limited deployment is the result of over a decade of work, it symbolizes the complex technical choices and fixations embraced by the world’s richest person in pursuit of fully autonomous vehicles.
Musk framed the idea of a driverless car as integral to the company’s future. This year it experienced a severe decline but he vowed to rapidly expand the Robotaxi service. Nonetheless, this week’s rocky launch suggests Tesla grapples with the technical hurdles that have drawn scrutiny from regulators.
The Robotaxi pilot involved around 10 cars navigating a confined area in Austin, with safety drivers present in the front seats. Additional limitations included restrictions during adverse weather and at nighttime. Influencer rides were priced at $4.20 each, mirroring Musk’s penchant for cannabis-related memes.
“Tesla’s autonomous driving can be deployed in approved locations. There’s no need for extensive mapping or specialized equipment,” the official Tesla account tweeted on launch day. “It just works.”
However, footage from at least 11 rides indicated that the trial did not unfold as flawlessly as Tesla’s promotional materials suggested. In one instance, the Robotaxi failed to make a left turn, veering into oncoming traffic instead, and resolved the issue by driving along a double yellow line. Other clips showed the vehicle allegedly exceeding speed limits.
This footage caught the NHTSA’s attention, with the agency stating they were aware of the incidents and had reached out to Tesla for more details.
Meanwhile, Musk retweeted a pro-Tesla influencer praising the service amidst technical failures and ongoing regulatory inquiries. One tweet shared by Musk featured a video showing a Robotaxi halting for a peacock crossing the road, while another urged followers to “ignore the media.”
“Lidar is lame.”
Musk has long maintained that reliance solely on cameras for autonomous vehicles is the key to true self-driving capabilities. Tesla’s consumer models feature what are termed “Autopilot” and “Fully Autonomous” capabilities, enabling hands-free driving on highways. These systems are supported by numerous external cameras for navigation, maneuvering, and stopping. The Robotaxis use similar software while depending entirely on cameras.
This camera-centric approach starkly contrasts with other self-driving tech firms like Waymo and Zoox, which utilize a combination of cameras and sensors, including radars and lidars. For instance, Waymo’s latest driverless vehicles are equipped with about 40 cameras and sensors, while Tesla’s advanced model for fully autonomous driving employs around 8 cameras. Bloomberg analysis. Lidar and radar are beneficial for detecting obstacles under poor weather and lighting conditions.
Despite lidar’s advantages, Musk argues that Tesla operates without it. “Lidar is lame,” he declared during Tesla’s Autonomy Day in 2019. “Using it in a car is foolish. It’s costly and unnecessary.”
According to Bloomberg, Lidar systems can cost around $12,000 each, whereas cameras are typically much more affordable. Musk contends that camera-only technology mirrors how humans navigate using their vision.
Tesla Faces Lawsuits and Investigations Over Full Self-Driving Mode
Musk’s claims regarding camera-only technology have placed Tesla under scrutiny, particularly following a fatal accident involving drivers using its fully autonomous driving features. The company is currently embroiled in various government investigations and civil lawsuits, asserting that fully autonomous driving suffers from weather-related issues like sun glare, fog, dust, and darkness. There are reports of at least 736 accidents and 17 fatalities linked to this technology. Analysis by the Washington Post.
“Tesla maintains an almost obsessive view of running the system solely on cameras, despite the consensus among experts in the field,” commented Brett Schreiber, a lawyer representing several victims of Tesla’s autopilot failures.
“Anyone following collision avoidance technology since the ’90s understands that radar, lidar, and cameras are the optimal trio.”
Schreiber expressed little surprise at Tesla’s Robotaxi’s shaky development in Austin.
“The real tragedy here is that people continue to be harmed and killed due to this technology,” he said. “And this highlights issues like, ‘Look how cute it is that a car can’t even make a left turn.’
Tesla did not respond to inquiries regarding the ongoing lawsuits, investigations, and crash incidents related to its fully autonomous driving capabilities.
Tesla’s Tactics vs. Waymo’s Approach
The contrast between Waymo’s method of launching commercial autonomous driving services in densely populated cities and Tesla’s approach extends beyond discussions about lidar versus cameras. Waymo is often seen as a frontrunner in the U.S. autonomous vehicle landscape, which has seen its competitors sharply reduced.
There are numerous reasons Waymo has outlasted many of its rivals. Historically, the Google subsidiary dedicated extensive time to mapping urban areas and rigorously testing vehicles prior to launch. For example, in San Francisco, where Waymo first implemented a completely autonomous commercial service, the company had begun mapping and testing as early as 2021.
Initiated as part of Google’s X Research Lab in 2009, Waymo also encountered challenges with self-driving cars despite its cautious, step-by-step city-by-city rollout. Earlier this year, Waymo was compelled to recall over 1,200 vehicles due to software problems causing collisions with roadside objects, gates, and other barriers. Additionally, the NHTSA launched an investigation last year after receiving 22 reports of Waymo vehicles demonstrating erratic behaviors or violating traffic laws.
In contrast, Tesla is still in the trial phase with its service, yet the Robotaxi launch in Austin marks the first time the automaker has deployed its fully autonomous driving technology in real-world conditions. There has been no information disclosed regarding the duration or extent of mapping or testing this technology in Austin.
This launch evokes memories of Uber’s initial attempt at self-driving vehicle ride-sharing services in 2016, which was also conducted without the necessary approval from California regulators. On the very first day of their pilot project in San Francisco, Uber vehicles reportedly ran a red light. They were forced to suspend the service just a week later after the DMV revoked their registration. At the time, an autonomous driving executive at Uber had urged engineers to expedite the process.
Faced with a lawsuit from Waymo regarding its self-driving operations and struggling to stay competitive, Uber sold its autonomous driving division in 2020.
Like Uber, Tesla also did not seek permission to operate its Robotaxi Service in Austin, as Texas has no existing permit process, which is not expected to be established until September.
At this time, it remains uncertain how frequently Tesla plans to deploy its Robotaxi service behind the scenes, but it’s clear that automakers are under pressure to meet deadlines set by Musk.
With the introduction of Robotaxis, Musk has claimed that Teslas will achieve full automation since at least 2016, and he may be approaching the deadlines he has postponed several times over the last decade.
Source: www.theguardian.com
