The December 2022 issue of IEEE Spectrum is here!

Close bar

Lane Keeping

Systems for keeping inside the lines are growing up, but they’re still not perfect

2 min read
Lane Keeping
Illustration: mckibillo

Drifting across the desert for a music video: cool. Drifting around a test track in an autonomous car: also cool. Drifting out of your lane on the highway: not cool. Carmakers have been warning drivers not to leave their lanes since 2001, with subtle hints such as audible beeps or vibrating steering wheels. These early systems used cameras to track lane lines painted on the road. A decade ago, Toyota’s first lane-keeping system took over the steering wheel and nudged wayward cars back into line when the driver would not. But because these systems relied on cameras and early image-processing algorithms, they worked only where the lines were clear and in good visibility.

Today’s lane-keeping systems are evolving. The Volkswagen Touareg can track a single lane stripe, and the company claims that the car can track lanes in the dark and in the fog. It manages the feat with a single camera, unlike some systems that use stereoscopic cameras or include information from a radar or other additional sensor.

That choice reveals some sophistication on the part of the image-processing software. Stereo vision would give the car’s processors depth perception to a certain distance, and radar might help it work when the sun is low, blinding the camera. Yet Volkswagen claims its single lane-keeping camera has, enough range for highway driving. It does not, however, work below 65 kilometers per hour (40 miles per hour), which may be a strategic decision to prevent drivers from trying to use the system in more dangerous and complex urban settings. Local roads have less-continuous lane markings than highways and are more likely to have obstructions blocking the car’s field of view and nearby obstacles. Bringing control to lower-speed lane-keeping systems would also require exerting more torque on the steering wheel than highway-speed lane-keeping systems do. The company would not provide further details on how its image-processing system works, but it is a safe bet that it was developed in conjunction with a third-party supplier and customized for Volkswagen.

Image processing is just the start, however: Senior research scientist Eddy Llaneras, at the Virginia Tech Transportation Institute, in Blacksburg, Va., says that a more interesting technical problem is the transition from warning to active control. It also defines the major divide between existing commercial systems. Some are just lane-departure warning systems; the best also nudge the wheel or brakes to help keep a car in its lane. The latter rely on already-existing stability control systems that prevent drivers from overcorrecting, for example, and which are part of today’s lane-keeping assistants.

Lane keeping requires the crucial latitudinal axis of car motion. Werner Huber, BMW’s head of driver assistance and perception, told IEEE Spectrum last year that “the big jump is now to take over control of the car in the longitudinal and latitudinal direction.” For cars that already have advanced cruise control, the temptation to let the car handle both axes is so great that carmakers, including BMW and Volkswagen, deactivate their driver assistance systems if the driver’s hands leave the wheel for more than a few seconds. Someday, the opposite may be true: The car may squawk if the driver tries to take the wheel.

Car Manufacturers

Maker Lane-Departure Warning Lane Keeping

Automotive parts suppliers

The Conversation (0)

Scaling Back a Quantum Compute Glitch by 100,000x

A software-based solution to qubits’ cosmic ray problem

3 min read
Conceptual computer artwork of electronic circuitry contained within spheres against beams of light, representing how data may be controlled and stored in a quantum computer.
Mehau Kulyk/Science Source

The kind of quantum computers that IBM, Google and Amazon are building suffer catastrophic errors roughly once every 10 seconds due to cosmic rays from outer space. Now a new study reveals a way to reduce this error rate by nearly a half-million-fold to less than once per month.

Quantum computers theoretically can find answers to problems no regular computer might ever hope to solve. Their key ingredients, known as quantum bits, or qubits, are linked together by a quantum effect known as entanglement.

Keep Reading ↓Show less

Deep Learning Gets a Boost From New Reconfigurable Processor

The ReAAP processor allows AI to be faster, more efficient

2 min read
different colored beams of light shooting up

This article is part of our exclusive IEEE Journal Watch series in partnership with IEEE Xplore.

Deep learning is a critical computing approach that is pushing the boundaries of technology—crunching immense amounts of data and uncovering subtle patterns that humans could never discern on their own. But for optimal performance, deep learning algorithms need to be supported with the right software compiler and hardware combinations.In particular, reconfigurable processors, which allow for flexible use of hardware resources for computing as needed, are key.

Keep Reading ↓Show less

Solving Automotive Design Challenges With Simulation

Learn about low-frequency electromagnetic simulations and see a live demonstration of COMSOL Multiphysics software

1 min read

The development of new hybrid and battery electric vehicles introduces numerous design challenges. Many of these challenges are static or low-frequency electromagnetic by nature, as the devices involved in such designs are much smaller than the operating wavelength. Examples include sensors (such as MEMS sensors), transformers, and motors. Many of these challenges include multiple physics. For instance, sensors activated by acoustic energy as well as heat transfer in electric motors and power electronics combine low-frequency electromagnetic simulations with acoustic and heat transfer simulations, respectively.

Multiphysics simulation makes it possible to account for such phenomena in designs and can provide design engineers with the tools needed for developing products more effectively and optimizing device performance.

Keep Reading ↓Show less