MIT Breaks Autonomous Drone Speed Limits By Not Sweating Obstacles

February 12, 2018 by Will Sweatman 29 Comments

How does one go about programming a drone to fly itself through the real world to a location without crashing into something? This is a tough problem, made even tougher if you’re pushing speeds higher and high. But any article with “MIT” implies the problems being engineered are not trivial.

The folks over at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have put their considerable skill set to work in tackling this problem. And what they’ve come up with is (not surprisingly) quite clever: they’re embracing uncertainty.

Why Is Autonomous Navigation So Hard?

Suppose we task ourselves with building a robot that can insert a key into the ignition switch of a motor vehicle and start the engine, and could do so in roughly the same time-frame that a human could do — let’s say 10 seconds. It may not be an easy robot to create, but we can all agree that it is very doable. With foreknowledge of the coordinate information of the vehicle’s ignition switch relative to our robotic arm, we can place the key in the switch with 100% accuracy. But what if we wanted our robot to succeed in any car with a standard ignition switch?

Now the location of the ignition switch will vary slightly (and not so slightly) for each model of car. That means we’re going to have to deal with this in real time and develop our coordinate system on the fly. This would not be too much of an issue if we could slow down a little. But keeping the process limited to 10 seconds is extremely difficult, perhaps impossible. At some point, the amount of environment information and computation becomes so large that the task becomes digitally unwieldy.

This problem is analogous to autonomous navigation. The environment is always changing, so we need sensors to constantly monitor the state of the drone and its immediate surroundings. If the obstacles become too great, it creates another problem that lies in computational abilities… there is just too much information to process. The only solution is to slow the drone down. NanoMap is a new modeling method that breaks the artificial speed limit normally imposed with on-the-fly environment mapping.

Continue reading “MIT Breaks Autonomous Drone Speed Limits By Not Sweating Obstacles” →

Living 3D Printer Filament

December 22, 2017 by James Hobson 15 Comments

This is more than a printing filament hack — closer to bleeding edge bio-engineering — but we can’t help but be fascinated by the prospect of 3D printing with filament that’s alive on a cellular level.

The team from MIT led by [Xuanhe Zhao] and [Timothy Lu] have programmed bacteria cells to respond to specific compounds. To demonstrate, they printed a temporary tattoo of a tree formed of the sturdy bacteria and a hydrogel ‘ink’ loaded with nutrients, that lights up over a few hours when adhered to skin swabbed with these specific stimuli.

So far, the team has been able to produce objects as large as several centimetres, capable of being adapted into active materials when printed and integrated as wearables, displays, sensors and more.

Continue reading “Living 3D Printer Filament” →

Printing Without Supports!

December 21, 2017 by James Hobson 19 Comments

If print supports have ever caused you grief, know that there’s an alternate printing method in the works. First: get yourself a vat of industrial gel in which to print.

Rapid Liquid Printing(RLP) is being developed in collaboration by Michigan-based company [Steelcase] and [Skylar Tibbits’] Self Assembly Lab at MIT. RLP is touting advantages over traditional 3D printing technology such as reduced print times, a higher quality print, and enabling larger scale prints — all without supports!

Working with rubber, plastic, or foam, the printing material is injected by nozzle into a basin of industrial gel. That gel suspends the print throughout the process without bonding to it and the finished product is simply lifted out of the gel and rinsed off. Shown off at the Design Miami event earlier this month, onlookers could pick up finished lampshades and tote bags after mere minutes.

Continue reading “Printing Without Supports!” →

MIT Is Building A Better 3D Printer

December 9, 2017 by Tom Nardi 79 Comments

Traditional desktop 3D printing technology has effectively hit a wall. The line between a $200 and a $1000 printer is blurrier now than ever before, and there’s a fairly prevalent argument in the community that you’d be better off upgrading two cheap printers and pocketing the change than buying a single high-end printer if the final results are going to be so similar.

The reason for this is simple: physics. Current printers have essentially hit the limits of how fast the gantry can move, how fast plastic filament can pushed through the extruder, and how fast that plastic can be melted. To move forward, we’re going to need to come up with something altogether different. Recently a team from MIT has taken the first steps down that path by unveiling a fundamental rethinking of 3D printing that specifically addresses the issues currently holding all our machines back, with a claimed 10-fold increase in performance over traditional printing methods.

MIT’s revolutionary laser-assisted hot end.

As anyone who’s pushed their 3D printer a bit too hard can tell you, the first thing that usually happens is the extruder begins to slip and grind the filament down. As the filament is ground down it starts depositing plastic on the hobbed gear, further reducing grip in the extruder and ultimately leading to under-extrusion or a complete print failure. To address this issue, MIT’s printer completely does away with the “pinch wheel” extruder design and replaces it with a screw mechanism that pulls special threaded filament down into the hot end. The vastly increased surface area between the filament and the extruder allows for much higher extrusion pressure.

An improved extruder doesn’t do any good if you can’t melt the incoming plastic fast enough to keep up with it, and to that end MIT has pulled out the really big guns. Between the extruder and traditional heater block, the filament passes through a gold-lined optical cavity where it is blasted with a pulse modulated 50 W laser. By closely matching the laser wavelength to the optical properties of the plastic, the beam is able to penetrate the filament and evenly bring it up to nearly the melting point. All without physically touching the filament and incurring frictional losses.

There are still technical challenges to face, but this research may well represent the shape of things to come for high-end printers. In other words, don’t expect a drop-in laser hot end replacement for your $200 printer anytime soon; the line is about to get blurry again.

Speeding up 3D printing is a popular topic lately, and for good reason. While 3D printing is still a long way off from challenging traditional manufacturing in most cases, it’s an outstanding tool for use during development and prototyping. The faster you can print, the faster you can iterate your design.

Thanks to [Maave] for the tip.

Continue reading “MIT Is Building A Better 3D Printer” →

AI Watches You Sleep; Knows When You Dream

August 29, 2017 by James Hobson 5 Comments

If you’ve never been a patient at a sleep laboratory, monitoring a person as they sleep is an involved process of wires, sensors, and discomfort. Seeking a better method, MIT researchers — led by [Dina Katabi] and in collaboration with Massachusetts General Hospital — have developed a device that can non-invasively identify the stages of sleep in a patient.

Approximately the size of a laptop and mounted on a wall near the patient, the device measures the minuscule changes in reflected low-power RF signals. The wireless signals are analyzed by a deep neural-network AI and predicts the various sleep stages — light, deep, and REM sleep — of the patient, negating the task of manually combing through the data. Despite the sensitivity of the device, it is able to filter out irrelevant motions and interference, focusing on the breathing and pulse of the patient.

What’s novel here isn’t so much the hardware as it is the processing methodology. The researchers use both convolutional and recurrent neural networks along with what they call an adversarial training regime:

Our training regime involves 3 players: the feature encoder (CNN-RNN), the sleep stage predictor, and the source discriminator. The encoder plays a cooperative game with the predictor to predict sleep stages, and a minimax game against the source discriminator. Our source discriminator deviates from the standard domain-adversarial discriminator in that it takes as input also the predicted distribution of sleep stages in addition to the encoded features. This dependence facilitates accounting for inherent correlations between stages and individuals, which cannot be removed without degrading the performance of the predictive task.

Anyone out there want to give this one a try at home? We’d love to see a HackRF and GNU Radio used to record RF data. The researchers compare the RF to WiFi so repurposing a 2.4 GHz radio to send out repeating uniformed transmissions is a good place to start. Dump it into TensorFlow and report back.

Continue reading “AI Watches You Sleep; Knows When You Dream” →

Design And 3D Print Robots With Interactive Robogami

August 26, 2017 by Donald Papp 2 Comments

Internals of 3D printed “print and fold” robot. [Image source: MIT CSAIL]

Robot design traditionally separates the body geometry from the mechanics of the gait, but they both have a profound effect upon one another. What if you could play with both at once, and crank out useful prototypes cheaply using just about any old 3D printer? That’s where Interactive Robogami comes in. It’s a tool from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) that aims to let people design, simulate, and then build simple robots with a “3D print, then fold” approach. The idea behind the system is partly to take advantage of the rapid prototyping afforded by 3D printers, but mainly it’s to change how the design work is done.

To make a robot, the body geometry and limb design are all done and simulated in the Robogami tool, where different combinations can have a wild effect on locomotion. Once a design is chosen, the end result is a 3D printable flat pack which is then assembled into the final form with a power supply, Arduino, and servo motors.

A white paper is available online and a demonstration video is embedded below. It’s debatable whether these devices on their own qualify as “robots” since they have no sensors, but as a tool to quickly prototype robot body geometries and gaits it’s an excitingly clever idea.

Continue reading “Design And 3D Print Robots With Interactive Robogami” →

The PDP-1: The Machine That Started Hacker Culture

June 27, 2017 by Dan Maloney 32 Comments

One of my bucket list destinations is the Computer History Museum in Mountain View, California — I know, I aim high. I’d be chagrined to realize that my life has spanned a fair fraction of the Information Age, but I think I’d get a kick out of seeing the old machines, some of which I’ve actually laid hands on. But the machines I’d most like to see are the ones that predate me, and the ones that contributed to the birth of the hacker culture in which I and a lot of Hackaday regulars came of age.

If you were to trace hacker culture back to its beginning, chances are pretty good that the machine you’d find at the root of it all is the Digital Equipment Corporation’s PDP-1. That’s a tall claim for a machine that was introduced in 1959 and only sold 53 units, compared to contemporary offerings from IBM that sold tens of thousands of units. And it’s true that the leading edge of the explosion of digital computing in the late 50s and early 60s was mainly occupied by “big iron” machines, and that mainframes did a lot to establish the foundations for all the advances that were to come.

Continue reading “The PDP-1: The Machine That Started Hacker Culture” →