Atlas robot jumps over a gap

Boston Dynamics Atlas Dynamic Duo Tackles Obstacle Course

Historically, the capabilities of real world humanoid robots have trailed far behind their TV and movie counterparts. But roboticists kept pushing state of the art forward, and Boston Dynamics just shared a progress report: their research platform Atlas can now complete a two-robot parkour routine.

Watching the minute-long routine on YouTube (embedded after the break) shows movements more demanding than their dance to the song “Do You Love Me?  And according to Boston Dynamics, this new capability is actually even more impressive than it looks. Unlike earlier demonstrations, this routine used fewer preprogrammed motions that made up earlier dance performances. Atlas now makes more use of its onboard sensors to perceive its environment, and more of its onboard computing power to decide how to best move through the world on a case-by-case basis. It also needed to string individual actions together in a continuous sequence, something it had trouble doing earlier.

Such advances are hard to tell from a robot demonstration video, which are frequently edited and curated to show highlighted success and skip all the (many, many) fails along the way. Certainly Boston Dynamics did so themselves before, but this time it is accompanied by almost six minutes worth of behind-the-scenes footage. (Also after the break.) We see the robot stumbling as it learned, and the humans working to put them back on their feet.

Humanoid robot evolution has not always gone smoothly (sometimes entertainingly so) but Atlas is leaps and bounds over its predecessors like Honda Asimo. Such research finds its way to less humanoid looking robots like the Stretch. And who knows, maybe one day real robots will be like their TV and movie counterparts that have, for so long, been played by humans inside costumes.

Continue reading “Boston Dynamics Atlas Dynamic Duo Tackles Obstacle Course”

Robot Pet Is A Chip Off The Old Logic Block

When [Ezra Thomas] needed inspiration for his senior design project, he only needed to look as far as his own robot. Built during his high school years from the classic 1979 Frank DaCosta book “How to Build Your Own Working Robot Pet”, [Ezra] had learned the hard way the many limitations and complexities of the wire wrapped 74xx series logic chips surrounding its 8085 processor.

[Ezra] embarked on a quest to recreate the monstrosity in miniature, calling it Pet on a Chip. Using a modern FPGA chip allows the electronics to shrink by an order of magnitude and provides flexibility for future expansion. Implementing an 8 bit CPU on the amply sized FPGA left plenty of room for a VGA GPU, motor controller, serial UART, and more. Programming the CPU is handled by a custom assembler written in Python.

The results? Twelve times less weight, thirteen times less power draw, better performance, and a lot of room for growth. [Ezra] hints at an I2C bus expansion as well as a higher level programming language to make software development less of a hurdle.

The Pet On A Chip is a wonderfully engineered project and we hope that we’ll be seeing more such from [Ezra] as time goes by. Watch his Pet On A Chip in action in the video below the break.

If [Ezra]’s FPGA escapades have you wondering how to get started, you can check out this introduction to FPGA from the 2019 Hackaday Superconference. And if you have your own FPGA creation to share, please let us know via the Tip Line!

Continue reading “Robot Pet Is A Chip Off The Old Logic Block”

Capstan Drive Is Pulling The Strings On This Dynamic Quadruped

When it comes to legged robots, it’s easy to think that the complexity and machining costs would keep these creatures far away from becoming anyone’s garage hobby. But, through a series of clever design choices, [Damian Lickindorf] has found a way to beat the odds and give life to Stanley, a low-cost, dynamic quadruped with some serious kick!

As if building a working legged robot weren’t already a tricky task, [Damian] has made some classy design choices to keep the price low and reduce fabrication complexity without sacrificing performance. Keeping up with the latest trend in Quasi-Direct Drive legged robots that started with the MIT Mini Cheetah, [Damian] constructed a small transmission with a gear reduction under 1:9. This choice slightly reduces the amount of heat produced by operating the motor at low-speeds with high torque without sacrificing too much control bandwidth (think: “leg responsiveness”).

Unlike the Cheetah, though, which uses a planetary gearbox, [Damian] opts for a capstan drive, a cable-driven transmission that’s both backlash free and backdriveable: two must-haves for force-sensitive dynamic legged robots. For legs, he’s opting for 2d machined FR4 (think: circuit board material). And for motors, he’s chosen a set of brushless motors with a large gap radius and driven by Moteus Drivers. The result is high fidelity, dynamic build that’s a fraction of the cost of some of the creatures we’re seeing emerge from academic research labs.

If you’re looking to feast your eyes on some action shots, look no further than [Damian’s] YouTube and Instagram presence. And if you’re looking to follow the project, have a look at the Hackaday.io project. While we’re eager to see the project continue to unfold, we’re thrilled by how far it’s come. In the meantime, be sure to take a look at one of the project’s inspirations: the Mjbots Quad A0.

Finally, since we’ve not seen capstan drives much on Hackaday, if you’re curious about these mechanisms and can get past the paywall, these two research papers might be a good place to dig deeper.

Continue reading “Capstan Drive Is Pulling The Strings On This Dynamic Quadruped”

Overhauling A Battle Bot

Where do old battle bots go to die? Well the great parts-bin in the sky corner of the workshop, where they await disassembly and use in other projects. But once in a while, if a battle bot is really lucky, they get pulled out again and put back into working order. So is the story [Charles] is telling about Overhaul 1, a hulk of a robot who was last see in fighting shape during the 2015 season of the show.

Having been succeeded by newer designs (Overhaul 2 and Overhaul 3), it’s a surprise to see some work being poured into these old bones. It didn’t escape the parts bin unscathed, having lost it’s wheels to another design called sadbot. What’s in place now are “shuffle drive pods”, a cam-based system that kind of crawls the robot along. They’re fun to watch in action in the video after the break, just make sure to turn your volume way down first. It’s no wonder [Charles] plans to replace them with newly-designed wheel modules.

In the heat of a match these things take a lot of damage, and the frame of Overhaul 1 was still twisted and mangled. A hydraulic tire jack is the tool of choice as the damage was caused externally and needed to be pushed out from the inside. As a testament to how these things are built, any old jack just won’t do and a 20-ton unit was acquired for the purpose. A set of prongs on the front (called pontoons) was also bent inward and required a chain and a come-along to pull them out.

The nice thing about revisiting projects years later is that technology tends to move forward. We can imagine that the design work [Charles] has in progress for a new set of wheel modules is much easier, and the parts (motors, drivers, batteries, etc) of a much higher quality than when first built over half a decade ago. This is the first installment in the overhaul of Overhaul series, which we’ll be keeping an eye on.

Need to sate your appetite for how to build indestructible robots? Check out how the indestructible wheels for the “Copperhead” bot are fabricated!

Continue reading “Overhauling A Battle Bot”

Ostrich Robot Machine-Learns Itself To 5K

Ever since humanity has grasped the idea of a robot, we’ve wanted to imagine them into walking humanoid form. But making a robot walk like a human is not an easy task, and even the best of them end up with the somewhat shuffling gait of a Honda Asimo rather than the graceful poise of a balerina. Only in recent years have walking robots appeared to come of age, and then not by mimicking the human gait but something more akin to a bird.

We’ve seen it in the Boston Dynamics models, and also now in a self-balancing two-legged robot developed at Oregon State University that has demonstrated its abilities by completing an unaided 5 km run having used its machine learning skills to teach itself to run from scratch. It’s believed to be the first time a robot has achieved such a feat without first being programmed for the specific task.

The university’s PR piece envisages a time in which walking robots of this type have become commonplace, and when humans interact with them on a daily basis. We can certainly see that they could perform a huge number of autonomous outdoor tasks that perhaps a wheeled robot might find to be difficult, so maybe they have a bright future. Decide for yourself, after watching the video below the break.

Continue reading “Ostrich Robot Machine-Learns Itself To 5K”

Create Large Scale Domino Art With A Robot

Creating large domino art displays is a long and nerve-racking process, where bumping a single domino can mean starting from scratch. To automate the process of creating these displays, a team consisting of [Mark Rober], [John Luke], [Josh], and [Alex Baucom] built the Dominator, a robot capable of laying 100 000 dominos just over 24 hours. Video after the break.

[Mark Rober] had been toying with the idea for a few years, and the project finally for off the ground after [Mark] mentioned it in a talk he gave at the 2019 Bay Area Maker Faire. To pull it off, the team created an entire domino laying system, including an automated loading station, a precision indoor positioning system, and the robot itself. The robot is built around a frame of aluminum extrusions, riding on three omnidirectional wheels driven by precision servo motors. A large tray mounted to the front of the robot can hold and release 300 dominos at a time. The primary controller is a Raspberry Pi 4, which receives positioning information from a Marvelmind indoor positioning system and a downward-facing IR camera that looks for reflective markers on the floor. The loading system uses a conveyor system to feed the different colored dominos to an industrial Kuka robot that drops them down a grid of tubes that can hold multiple layers at once.

Continue reading “Create Large Scale Domino Art With A Robot”

Tabletop Basketball With Tentacles

Unlike football/soccer and foosball, basketball doesn’t really lend itself to being turned into a tabletop game quite that easily. [The Q] has found a way around that, employing tentacle mechanisms to create a two-player, basketball-like game.

Each player uses a pair of two-axis control sticks and a foot pedal to operate a cable-driven tentacle with a gripper on the end. These are two stage tentacles, meaning that the top and bottom halves are independently controlled. The tentacles consist of a series of laminated foam discs clued onto bicycle cable sleeves. The cables are open in the section they control, and operate in a push-pull arrangement. The spring-loaded grippers are operated by the foot pedals, with a single cable running down the center of the tentacle.

The game looks quite fun and challenging, and we can imagine it being even more entertaining with teams of two or three people operating each tentacle. Add a bit of alcohol to adult players, and it might become downright hilarious, although the mechanisms would need to be beefed up a bit to survive that level of punishment.

We suspect [The Q] read [Joshua Vasquez]’s incredibly detailed three-part guide on two-stage tentacle mechanisms. Combine that with his guide to cable mechanism math, and you’d be well-equipped to build your own. Continue reading “Tabletop Basketball With Tentacles”