Evaluation Engineering - 36

IGphotography/iStock/Getty Images Plus

TECHNOLOGY INSIGHTS

ROBOTIC VEHICLES ROLL INTO REALITY
By Ken Cormier, Managing Editor
Self-driving or autonomous vehicles,
gliding down the street without a
driver actively operating any of the vehicle's control systems, are quickly emerging from the mist of sci-fi dreams.
Relying on sensors such as radar,
LiDAR, sonar, GPS, odometry and inertial measurements to perceive their surroundings, robotic vehicles are currently
being tested and fine-tuned by companies such as Tesla, Waymo, and Uber.
According to the National Conference
of State Legislatures, at least 29 states
have enacted legislation related to autonomous vehicles.
In February of this year, Tesla CEO Elon
Musk said that he anticipates the completion of a fully autonomous vehicle by the
end of this year. In April, he predicted that
"probably two years from now, we'll make
a car with no steering wheels or pedals."
Fasten your seatbelt. It looks like we're
in for an interesting ride.

Ford shoos bugs from auto sensors
In testing the LiDAR sensors on its selfdriving cars, Ford has gone the extra mile
to make sure they can function consistently under realistic circumstances.
Some of the methods Ford has employed
include showering sensors with water,
dirt and dust, smearing cameras with
synthetic bird droppings, and enlisting
a "bug launcher" that shoots insects into
sensors at high speed.

EVALUATION ENGINEERING OCTOBER 2019

Test engineers took advantage of the
"tiara" on top of the car that houses the
cameras, LiDAR, and radar. The tiara was
used to funnel air through slots near the
camera lens which created an air curtain
that deflected incoming insects. Not satisfied with the results, engineers added a
system of nozzles near the lens to spray
washer fluid in tandem with software algorithms that would determine when a
cleaning was needed. Air flowing through
tiara slots were used as a drying agent.1

Science and Artificial Intelligence
Laboratory (CSAIL) and the Andrew
and Erna Viterbi Professor of Electrical
Engineering and Computer Science. "For
example, if we train an autonomous vehicle to drive in an urban setting such
as the streets of Cambridge, the system
should also be able to drive smoothly in
the woods, even if that is an environment
it has never seen before."2

Allowing self-driving cars to
make more human decisions

John Rich, operations head for Ford
autonomous vehicles, recently stated
that Ford autonomous vehicles will last
only four years. His prediction doesn't
mean the autos won't be sturdy enough
to log in high mileage-Rich thinks the
vehicles are going to be used more heavily than "normal" use. His reasoning is
that decreasing ownership of autos will
lead to consumers hailing autonomous
vehicles when needed. Such a scenario
might lead to the vehicles running 24/7,
and thus the accelerated mortality of the
robotic carriages. 3

MIT researchers have come up with an
enhanced end-to-end navigation system
designed to help driverless cars negotiate
roads they've never encountered before.
The system uses maps and visual data,
along with machine learning, to learn the
steering patterns of a human driver, and
later replicate the patterns in unfamiliar territory, using only data from video
camera feeds and a simple GPS-like map.
"With our system, you don't need to train
on every road beforehand," says first author Alexander Amini, an MIT graduate
student. "You can download a new map
for the car to navigate through roads it
has never seen before."
"Our objective is to achieve autonomous navigation that is robust for driving in new environments," adds co-author
Daniela Rus, director of the Computer

Ford exec.: Ford autonomous
vehicles will last only four years

REFERENCES

1. Ford Authority, "This is How Ford
Autonomous Vehicles Keep Sensors Clean,"
August 2019
2. MIT News, "Bringing human-like reasoning to
driverless car navigation," May 2019
3. Ford Authority, "Ford Autonomous Vehicles
Will Only Last Four Years," August 2019

Evaluation Engineering

Table of Contents for the Digital Edition of Evaluation Engineering