Want to Build a Better Bot? JetsonHacks Shows the WayDecember 10, 2015
Build your own robot. It’s a task that sounds daunting for most, easy for some.
Regardless of where you stand, you need to meet Jim Benson, a computer scientist turned robot enthusiast who hosts the JetsonHacks.com website and YouTube channel. Benson shows bot builders that they’re just are a few steps away from DIY heaven.
The JetsonHacks.com community focuses on what can be done with low-power, supercomputer-level GPUs, starting with the NVIDIA Jetson TK1 Developer Kit. With the recent introduction of the Jetson TX1 — capable of powerful tasks such as computer vision, machine learning and navigation — hardware and software developers will be salivating at the possibilities.
“For a robot designer, the Jetson TX1 offers two things that are a big deal — an off-the-shelf module and low power consumption,” Benson said.
Compact Design, with a Supercomputer at the Heart
During a career in geospatial intelligence, Benson developed ways to visualize and interact with digital maps. Intrigued by the power and abilities of GPUs in mobile phones and other devices, he could see computer power growing ever stronger.
Smartphones and tablets are much a richer computing environment than the desktop because they contain hardware sensors and interface devices such as multiple cameras, multiple radios, accelerometers and gyroscopes, touch screens and GPS systems.
“At the heart of these systems are devices like the Jetson,” Benson said. “The Jetson product is one of those inflection points in computer history where embedded and desktop merge.”
JetsonHacks.com provides actionable information for developing on Jetson, with step-by-step instructions, materials lists, and insight on dev kits, among other tips.
We dub Benson a “Jetson Champ” because, like eight other Champs around the world, he offers inventive solutions to engineering problems, guidance on tricky mechanical issues and tutorials that help people of all ages learn the skills to build professional-quality robots.
Showcasing lessons online means ideas in JetsonHacks.com may cross-pollinate with others, and lead to the next big thing. The network effect of Benson’s foray online, which began in 2014, surprised him. The developer appetite for all things robot is strong.
The more than 130 posts on the JetsonHacks website related to the Jetson dev kits include software installation, deep learning, hardware interfacing and robotics. There’s a 17-part series on building the JetsonBot, an inexpensive vision robot based on the ROS TurtleBot powered by the Jetson TK1.
The JetsonHacks YouTube channel offers about 100 Jetson dev kit-related videos. In little more than a year, the videos have had over 145K views, a number that is growing as much as 20 percent a month.
What’s more, the JetsonHacks Github code repository has 40 Jetson-related open source code projects available for download, as well as 25 “gists” for helping with software installation and projects.
As more schools add robotics to the curriculum and the interest in robots grows, so does demand for information from developers. With the super-powerful Jetson TX1 processer unleashed, Benson’s lessons are a springboard for what comes next.