Hydraulically-actuated robots are nothing new, but normally they come with a battery or external supply of some sort. This lifelike robotic lionfish developed by Cornell and the University of Pennsylvania researchers, however, has its own artificial circulatory that pumps synthetic ‘blood’ to help flap its fins and as the device’s power source itself. 

The trick is that the liquid is actually the cathode of a battery built into the fish, which powers its two hydraulic actuators, as well as the Arduino Uno control system. This integral battery—which would be analogous to blood in a real fish—gives it enough energy to operate untethered for 36 hours, though as it swims at 1.56 body lengths per minute, so it can use all the time it can get!

As James Pikul, a co-author on the study and researcher at Penn, told Gizmodo:

In our synthetic vascular system, the fluid stores chemical energy which we can use to power the fish robot. As the fluid is pumped through the fish robot, the moving fluid also causes the robot to move. The vascular system, therefore, is multifunctional. It is these multiple functions that allow the robot to maintain its dexterity while also having a long operational time.

You can also read more in IEEE Spectrum‘s article here.

You’ve seen self-balancing robots, where a pair of wheels suspend a mass above them in what’s known as an inverted pendulum configuration. As neat as they are, the “Augmented Arthropod” by Grzegorz Lochnicki and Nicolas Kubail Kalousdian puts a new spin on things. 

The structure for the build consists of three platforms separated on threaded rod and a couple of rather standard DC gear motors. Electronics include an Arduino Uno, a BNO055 IMU, and an L298N motor driver. 

Where things get a bit interesting, though, is that the mech is piloted by the movements of an insect placed inside a plastic case using two HC-SR04 ultrasonic sensors. 

Perhaps the most valuable part of the project write-up is the discussion about how it balances via PID, or proportional, integral, and derivative control. 

While the hoverboard craze has faded somewhat, the good news is that this means their powerful wheel motors can easily be found on online auction sites. Lukas Kaul took advantage of this component’s availability and created his own “HoverBot,” which as shown in Felix von Drigalski’s video below, acts as something in between a radio-controlled skateboarder and a rather large self-balancing bot.

The device is built around an Arduino Mega, which takes input from an RC receiver, along with a Bosch BNO055 IMU, and passes appropriate signals to the motors through an ODrive controller. 

The HoverBot is a bit unsteady at high speeds, requiring close operator supervision. However, it looks like a lot of fun, especially when attempting tricks—sometimes successfully—at a skate park.

Since its invention in 1974, Rubik’s cubes have been entertaining and frustrating those that choose to take on the challenge of aligning their shapes. More recently, however, people have been building algorithms and machinery to do it for them, including Mario Milanesio’s Arduino Rubik Solver, or ARS.

ARS, which was constructed with the help of Milanseio’s students, is comprised of several 3D-printed and laser-cut parts. The device utilizes a series of four stepper motors to rotate the cube, along with two more to pull the grippers back when needed. 

Solving is assisted by the ARS Studio software package, which lets users program in the existing color sequence. It then sends movement commands to an Arduino Uno over serial, which controls the motors via six A4988 Pololu drivers to complete the puzzle.

VR environments are meant to be immersive, but if you’ve ever thought what was missing is being actually pummeled by robotic fists, then James Bruton’s newest project could be just the thing. 

Bruton recently teamed up with students from Portsmouth University to build a robot that works in the real world, and coordinates its movements with a virtual setting displayed on the human’s headset.

The robot itself is controlled by an Arduino Mega, and features a differential (tank) drive with encoders for feedback. Shoulders can tilt from left to right, and the actual punching motion is handled by pneumatic actuators built from modified bicycle pumps. Robo-fists are covered by boxing gloves to keep humans relatively safe, and flesh-based competitors are given a small shield and sword-bat with which to fight back!

I worked on this project with final year degree students in Computer Games Technology at Portsmouth University CCI faculty. The robot hardware is controlled over a serial interface, the team built an VR game which controls the robot, so when you get hit in VR you get hit in real life! The robot is tracked back into VR with Vive trackers so it stays in sync.

If you have to do a lot of drawing on a whiteboard, you also have to clean it. Why not have a robot do this instead? That’s the idea behind Wipy, an Arduino Uno-based robot that uses magnets to stick it to the board, plus grippy wheels and motors to power it across your scribbles.

Wipy employs an array of IR sensors that enable it to act as a line follower, along with a time-of-flight (ToF) sensor to detect your hand on the board. While one might assume this sensing arrangement would prevent it from erasing your work-in-progress, it annoyingly allows it to start erasing immediately when you start drawing on the board. At least it has a cute LED face!

Did you ever get tired of cleaning the whiteboard? Have you ever wondered how much your life would improve if a robot could do this for you? You now have the chance to make this a reality with Wipy: the overly motivated whiteboard cleaner. Wipy will properly clean your embarrassingly bad drawings, and it will even do it with a cute smile. You don’t even need to activate it! It will just clean the board when you least expect it… Uhhh…*cough cough*…we, of course, mean: when you need it most!

– Our future friend will be able to stick to the board using magnets and is able to move through space using grippy wheels.

– It will be able to follow a line and erase it using a line-following sensor and a sponge.

– Wipy has the ability to measure the distance to your hand using a time-of-flight sensor.

– We will give Wipy a cute personality using a small OLED screen.

Do you like plants, but not so much the tending to and watering them? If that sounds like you, then you might be interested in your own CNC growing machine. The system—created by 15-year-old maker “daily3dprinting”—is controlled by an Arduino Uno, and uses a single stepper motor to pull a watering head into position based on hygrometer readings.

A relay is used to turn the grow light on at 6am and off at 8pm, and another to activate the unit’s water pump. A third relay is employed to power off the L298N stepper driver when not needed. 

The project took home second place in the math and engineering category at daily3dprinting’s high school science fair, and more info on the build is available in its write-up here.

Would you like a dog? Would you like a robot dog? If so, then this build by Michael Rigsby could be a great starting point. 

Rigbsy’s robotic pet features four servo-driven legs, with two-axis shoulder movement, as well as an articulated knee joint. As seen in the video below, it’s capable of picking itself up off the ground, and can then walk using a slow side-to-side gait.

An Arduino Uno uses the majority of its I/O pins to control the legs, and as of now, it travels forward with no directional control or sensor input. 

Instructions for the project, along code and 3D print files, are available in Rigsby’s write-up.

In a variety of robotic situations, you’ll need some sort of gripper. In this project, James Bruton attempts to create a force-controlled, three-fingered assembly using an Arduino Uno along with a trio of servos.

Instead of directly controlling the grip fingers, the 3D-printed device is held open with bungee cables. When it’s time to clamp everything down, the servos wind up the cables attached to the inside of the fingers, similar to how human tendons work. 

To correlate servo inputs to grip force, he uses a series of springs to allow some amount of compliance, as well as flex sensors attached to the fingers themselves to measure the resulting positions. Arduino code for the build is available here.

Rubik’s Cubes seem to have been most popular in the 1980s, but never really went away. As such, if you have one lying around your house unsolved, why not ‘simply’ construct a machine to do this for you? 

One possibility is the Q-Bot, outlined here. While it won’t break any world records, it’s a solid-looking assembly that appears to be relatively easy to build.

The Q-Bot features six NEMA 17 stepper motors, four of which turn one face at a time. When needed, the other two use timing belts to alternatively pull opposed stepper motor pairs back, allowing the other two to rotate the entire assembly. An Arduino Mega is utilized to control the steppers via a custom shield, with a computer running the Kociemba’s Algorithm.