Google Patents Driverless Car Tech for Sliding In and Out of Tight Spots
The U.S. Patent Office has granted Google a patent that covers the way in which cars could transfer from human-driver mode to autonomous-driving mode. Google has already conducted many driverless-car experiments, claiming it's logged 200,000 miles of driver-free travel. However, it's unclear how extensively this technology will be used in the future.
12/16/11 12:19 PM PT
Google has been awarded a United States patent for what, in legalese, is "transitioning a mixed-mode vehicle to autonomous mode" -- essentially a self-driving car.
To be more precise, U.S. Patent 8,078,349 covers a car being driven by a human that transitions to being autonomously driven.
This transition may include stopping a vehicle on a predefined "landing strip" and detecting a reference indicator which tells the vehicle its exact position and provides an URL that links to an instruction to go autonomous.
What the Google Patent Covers
The patent outlines various methods and devices for switching a mixed-mode vehicle from being driven by a human into autonomous mode.
One method might be the use of a predefined landing strip. Others might be placing a reference line for the vehicle to follow, programming a predefined route into a vehicle, or having a computer control an autonomous vehicle that follows a route based on information stored in the computer.
However, Google might opt for the predefined landing strip method, on which the patent primarily focuses.
Equipment for the Landing Strip Method
An autonomous vehicle may be equipped with a landing sensor to detect the landing strip as well as a computer system to process the data.
A reference indicator may be embedded on the ground within the landing strip or mounted in line of sight of a sensor such as a camera on the vehicle. It may be a QR code, a one-dimensional bar code or a radio tag.
A QR, or Quick Response, code is a two-dimensional bar code originally designed for the automotive industry. It consists of black patterns on a white background in a square. The information encoded in the code can be any kind of data.
The autonomous vehicle may be equipped with both optical and radio sensors so it can detect more than one type of reference indicator. These might double up as reference sensors.
Workings of the Landing Strip Method
A driver will have to stop a mixed-mode autonomous vehicle on the landing strip to transition it to autonomous drive mode. The vehicle's sensors would detect the landing strip and reference indicator.
The reference indicator may contain reference data -- an URL, a physical address or global coordinates, for instance -- which will be decoded by the computer in the vehicle.
It could also provide data related to autonomous driving instructions. These can either be stored in a database in the vehicle's memory or be transmitted wirelessly to the vehicle.
Remarks on the Technology
It's not likely that public roads will be configured with sensors for autonomous vehicles any time soon because "it's generally believed to be too expensive to put the sensors in the road," Rob Enderle, principal analyst at the Enderle Group, told TechNewsWorld. "Google has been testing visually oriented technology and shouldn't need a heavy sensor base."
Sensor technology "may be used where there's heavy traffic or where tolerances are tight, like say in a car dealership or manufacturing plant," Enderle speculated.
Google "may be able to answer" some additional questions raised about the technology "more meaningfully in the new year," Jay Nancarrow, a spokesperson for the company, told TechNewsWorld. However, "we aren't going to speculate about specific applications of the technology at this time," he added.
Google's driverless cars have "driven roughly 200,000 autonomous miles" over the past few years, Nancarrow said.
The state of Nevada has already amended its laws to allow autonomous vehicles on its roads.
Other Autonomous Experiments
Researchers at the China National University of Defense Technology equipped a Hongqi HQ3 with video cameras and radar centers controlled by a command center in the trunk and sent it on a three-and-a-half-hour drive in highway traffic from Changsha to Wushan in July.
Separately, a group of Italian engineers was reported in July to be planning to send two driverless electric vehicles equipped with four laser scanners and seven video cameras and technicians behind the wheel on an 8,000-mile journey from Italy to China.