Three Startups Find New Ways to Help the Planet and People

by Tonie Hansen

GPU computing has helped create amazing advances in self-driving cars, virtual reality and artificial intelligence. Industries the world over are being transformed.

Among my favorite uses for GPU computing: helping the planet. The three startups below are taking advantage of the technology for a purpose, with an eye to reducing energy use, cutting waste and keeping people safe.

Bright Lights, Less Energy

Public lighting to illuminate streets, parking lots and other outdoor areas consumes tons of energy. But most people are unaware of the financial and environmental costs that come with it.

Smart-I, of Italy, has found a way to significantly reduce energy consumption with its AI-powered system SmartEye. Powered by NVIDIA Tegra Processors, SmartEye uses optical sensors to monitor road conditions in real time so it can adjust lighting levels and increase road safety, while providing the highest energy savings for outdoor lighting.

SmartEye’s AI brain identifies optimal lighting levels for varying weather conditions, driving visibility and traffic flow. In addition to reducing energy use, SmartEye provides interactive services, such as advanced monitoring to increase city security.

Robots Take on Recycling

In most recycling facilities, workers sort through waste streaming down a conveyor, separating valuable materials from garbage. This process is unpleasant, potentially dangerous and costly because of the extensive labor required.

AMP Robotics, of Denver, Colorado, seeks to address these challenges — and chop recycling costs in half — by automating the sorting process. Its system relies on GPU-powered machine learning to teach robots how to recognize different kinds of objects: from opaque and shiny plastic, to popular logos like those of Pepsi and Coca-Cola.

Thousands of pictures are shown to the system to help it learn to identify distinguishing features in the waste stream. AMP Robotics aims to create robotic devices that are easy to install and durable, efficiently sorting through recycling while keeping human limbs and hands safe.

Faster, Better Weather Forecasts

People love to complain about inaccurate weather forecasts. But dealing with the inconvenience of a drizzly picnic is a far cry from slow or inaccurate warnings for hurricanes or typhoons, which can lead to destruction of infrastructure, property and lives.

TempoQuest, based in Las Vegas, is developing environmental and weather forecasting software to create faster, more accurate projections. Its software, which runs on NVIDIA Tesla accelerators, speeds up the Weather Research and Forecasting model — the world’s most used weather prediction system.

Tempo Quest accelerates the speed of forecast processing, improves forecast accuracy by a factor of three and increases the frequency of forecast distribution, giving people a chance to seek proper shelter and prepare for worst case scenarios.

Corporate Sustainability

From modeling climate change to enhancing data center efficiency, GPUs are used extensively for environmental purposes. Find out more, and learn about NVIDIA’s environmental priorities, in our 2016 Corporate Sustainability Report.