Phantom Auto Developing a Remote Control Option for Autonomous Vehicles

Home > News > Content

【Summary】Phantom Auto hopes to relieve some of the ‘ride-anxiety’ associated with self-driving cars — a human safety driver, with the ability to remotely control a self-driving vehicle when needed.

Original   Eric Walz  ·  Dec 06, 2017 3:13 PM PT
author: Eric Walz   

MOUNTAIN VIEW, Calif., —  As autonomous cars edge closer to reality, the idea of being a passenger in a car with no driver is frightening to many people. A new Silicon Valley startup has emerged from stealth mode to address this problem. Phantom Auto hopes to relieve some of the ‘ride-anxiety' associated with self-driving cars — a human safety driver on standby, with the ability to remotely control a self-driving vehicle when needed.

Phantom Auto is developing technology that allows a human driver to take control of a car remotely over a secure 4G network, and steer around construction zones, other vehicles, emergency vehicles, or pedestrians. The technology can even be used in bad weather, to help navigate a vehicle in heavy rain or fog.


Phantom Auto software engineer Ben Shukman operates a car remotely.

The company was founded by CEO Shai Magzimof, along with Ohad Dvir who serves as the company's COO, and Andrew Gryaznov, the company's CTO. Phantom's U.S. headquarters is located in Mountain View, California, with another office in Tel Aviv, Israel.

Prior to launching Phantom Auto, Magzimof was the former CEO and co-founder of Nextpeer - a mobile gaming social platform before starting Nextpeer at age 20, Shai started his first company with a childhood friend while participating in Y-Combinator. This company later became Cue, which was acquired by Apple.

Dvir served as the Project Manager for Military Autonomous Vehicle Division at Israel Aerospace Industry and Gryaznov has experience in artificial intelligence, telecommunications, and cloud computing.

How Phantom Auto's Technology Works

Using multiple monitors and a driving simulator that resembles a car racing video game complete with a steering wheel, brake pedal, accelerator, a remote operator can "see" exactly what the cars camera are viewing in real-time and allows a human operator to take manual control of the vehicle when required. This can be useful in certain situation where the vehicle's software encounters a situation that it cannot resolve with its current hardware or machine-learning algorithms.

San Francisco based self-driving startup Cruise Automation encountered a similar problem at a media event last week. Cruise was conducting a ride along with members of the media to demonstrate the self-driving capabilities of its fleet of Chevy Bolt EVs on the streets of San Francisco. However, one of the vehicle's encountered a situation where one of its Chevy Bolt test vehicles braked and remained stopped behind a food truck blocking the lane ahead, with no attempt by the vehicle's on board-software to steer around the obstacle.

A human safety driver was forced to take manual control of the vehicle to steer around the stopped food truck. While the incident did not pose a major safety risk, it shows some of the inherent difficulties a self-driving car faces when navigating on public roads. In the current stage of development, there are many situations where human intervention is necessary for a autonomous car.

Phantom Auto's solution can prevent this from happening. More importantly, it gives riders peace of mind, knowing that a human safety operator is monitoring the vehicle and is ready to take over if necessary.

Prev                  Next
Writer's other posts
    . Related Content