As far as controlling robots goes, makers today are spoilt for choice. WiFi and Bluetooth enabled microcontrollers are a dime a dozen, and integration with smartphone apps is a cinch. Despite this, the old methods still hold sway, as [Igor Fonseca] demonstrates with a simple Arduino bot.
It’s a classic build, using a tracked chassis with a pair of motors providing propulsion and skid steering. The motors are controlled by an L298N H-bridge board, with power courtesy of a trio of 18650 batteries. An Arduino Uno acts as the brains of the operation. Control is via a Playstation 2 controller, in this case a 2.4 GHz third party version. This allows the robot to be controlled wirelessly, with the decoding handled by [Bill Porter]’s useful Arduino library.
It’s a cheap approach to building a remote-controlled bot, and one that would be a great way to teach interested children about how to work with embedded systems. We’ve featured a similar build before, too. Video after the break.
India’s Chandrayaan-2 mission to the Moon was, in a word, ambitious. Lifting off from the Satish Dhawan Space Centre on July 22nd, the mission hoped to simultaneously deliver an orbiter, lander, and rover to our nearest celestial neighbor. The launch and flight to the Moon went off without a hitch, and while there were certainly some tense moments, the spacecraft ultimately put itself into a stable lunar orbit and released the free-flying lander so it could set off on its independent mission.
Unfortunately, just seconds before the Vikram lander touched down, an anomaly occurred. At this point the Indian Space Research Organisation (ISRO) still doesn’t know exactly what happened, but based on the live telemetry stream from the lander, some have theorized the craft started tumbling or otherwise became unstable between three and four kilometers above the surface.
Telemetry indicates a suboptimal landing orientation
In fact, for a brief moment the telemetry display actually showed the Vikram lander completely inverted, with engines seemingly accelerating the spacecraft towards the surface of the Moon. It’s unclear whether this was an accurate depiction of the lander’s orientation in the final moments before impact or a glitch in the real-time display, but it’s certainly not what you want to see when your craft is just seconds away from touchdown.
But for Chandrayaan-2, the story doesn’t end here. The bulk of the mission’s scientific goals were always to be accomplished by the orbiter itself. There were of course a number of scientific payloads aboard the Vikram lander, and even the Pragyan rover that it was carrying down to the surface, but they were always secondary objectives at best. The ISRO was well aware of the difficulties involved in making a soft landing on the Moon, and planned their mission objectives accordingly.
Rather than feel sorrow over the presumed destruction of Vikram and Pragyan, let’s take a look at the scientific hardware aboard the Chandrayaan-2 orbiter, and the long mission that still lies ahead of it.
These days, it can feel like a project doesn’t exist unless you’ve posted a video on the Internet about it. [mingul] was in the process of producing his own videos, but found having to repeatedly move and set up the camera tiring. Naturally, a completely overkill eight-axis motion control robot was the solution. Video embedded below the break.
The scale of the build is something to behold. With 4.5 m travel on the X-axis, 6.5 m on the Y, and 2.1 m on the Z, it’s capable of traversing the full length of [mingul]’s workshop. Tilt, pan, and roll axes all feature 540 degrees of rotation, and there’s motors to control zoom and focus on the camera, too. Through software like Dragonframe, it’s possible to program complicated camera moves, and techniques like the classic dolly zoom are a cinch with such a versatile rig. It’s also possible to control the movement in real-time with a wireless Xbox controller.
[mingul] reports the build took a full three months of CNC machining, 3D printing and assembly. It’s a big step above a simple motorized camera slider, but we all have to start somewhere.
Pong is a classic from the very dawn of the video game era. Recreating it remains a popular exercise for those new to coding. However, its simple logic makes this game particularly suited to an all-hardware build; something which [Glen] tackles with aplomb.
Not content to take the easy way out, [Glen] went for a particularly hardcore method of construction. The game uses absolutely zero integrated circuits in its construction. Instead, it relies upon the services of 431 bipolar transistors, 6 JFETs and 826 diodes. Everything is laced together on protoboard, connected with a neatly organised nest of colored wires. Schematics are available for the curious.
It’s a full featured build, too. Video output is in color, scores are displayed at the top of the screen, and there’s even stereo panning for the sound effects. It just goes to show what some humble components can do when put to work in the right way. We’ve seen some of [Glen]’s work before too, for example in this op-amp bouncing ball device. Video after the break.
Sometimes you need something really flashy to complete an outfit. Whether it’s a sparkly pair of earrings or a stylish necklace, accessories are key to competing on the fashion battlefield. For those who want to bring some serious firepower, [p3nguin’s] laser crown might be just what the doctor ordered.
At the outset, we should state the crown only uses lasers in its construction, for cutting felt and acrylic. The light source is a Neopixel ring from Adafruit, capable of bringing the vibrant colors without risk of eye damage. The ring is then assembled with a series of snap-together acrylic parts and a felt cap, with slots for hair pins to keep the crown in place on the wearer’s head. A Trinket drives the show, with a LiPo battery used as a lightweight power supply.
It’s a nice build that’s sure to draw plenty of attention. We see some great wearables around these parts; this EL jacket is a particular favorite. Video after the break.
These days, working with a display in software is fairly easy. Thanks to the convenience of the modern OS, we’re blessed with graphical user interfaces, where things such as buttons and windows and text are all taken care of for us. Of course, once you start to wander off the beaten track, particularly in embedded systems with no GUI, things can get a little more difficult. For these situations, [JSBattista] wrote some code to blast text directly to the Linux framebuffer.
It’s a project borne out of necessity. Working with a Raspberry Pi with no X server, it was found that the console text size made it difficult to display data. By writing directly to the framebuffer, it would be possible to display text of a larger size without having to implement a full GUI, and overheads could be kept to a minimum.
Working in this manner comes with some limitations. Glyphs are taken from an array in bitmap format, rather than font files. In this case, a font akin to that of the Alien sentry gun interface was chosen, for an attractive sci-fi look. Lowercase characters are currently unimplemented. Testing thus far has been on Raspberry Pi and Beaglebone non-GUI systems, with performance varying depending on platform.
It’s a project we suspect might prove useful to the developers of lightweight embedded systems. It’s something that may take some tweaking and experimentation to implement, but the hacker set rarely shy away from a challenge. If you’re eager to get down and dirty with some heavy programming, this tutorial on Linux graphics will help.
When starting a new job, learning coworkers names can be a daunting task. Getting this right is key to forming strong professional relationships. [Ahad] noted that [Marcos] was struggling with this, so built the Name Stone to help.
The Name Stone consists of some powerful hardware, wrapped up in a 3D printed case reminiscent of the Eye of Agamotto from Doctor Strange. Inside, there’s a Jetson Nano – an excellent platform for any project built around machine learning tasks. This is combined with a microphone and camera to collect data from the environment.
[Ahad] then went about training neural networks to help with basic identification tasks. Video was taken of the coworkers, then the frames used to train a convolutional neural network using PyTorch. Similarly, a series of audio clips were used to again train a network to identify individuals through the sound of their voice, using MFCC techniques. Upon activating the stone, the device will capture an image or a short sound clip, and process the data to identify the target coworker and remind [Marcos] of their name.
