One of the challenges of many walking robot designs is the fact that they draw current just to stay upright. This was exactly the case for one [James Bruton]’s quadruped robots, where the knee motors were getting too hot to touch. Adding springs to take some of the load is not as simple it might seem, so [James] created a bungee assisted cam mechanism to do the job.
For a normal spring-loaded lever, force is proportional to how much the spring is stretched, which will require the actuators to draw more and more current as it lifts the leg higher. For the spring force to remain constant throughout the range of motion, the length of the lever arm must become continuously shorter as the knee is bent. [James] did this by stretching a bungee cord around a cam. The added bulk of the cam does however cause the knees to knock into each other in some scenarios, but [James] plans to adjust the robot’s gait to avoid this. He didn’t get around to actually measuring the current draw reduction, but the motor temperature has dropped significantly, only being slightly warm after a test run.
The ability to move in any direction and turn on the spot is a helpful feature on robots that operate indoors around other objects. [James Bruton] demonstrated one possible solution in the form of a robot chassis that can move in any direction with three ball-shaped wheels.
The video after the break is part two of this series. Part one covered the ball wheels themselves, consisting of a pair of half-spheres that can rotate independently with a small roller in the center of each and a driven shaft through the center of the sphere. Three of these are arranged at 120° intervals around the center of the robot, with the main shafts driven by geared DC motors using belts. To move in a straight line some basic trigonometry is used to calculate the required relative speed of each wheel. An Arduino Mega is used to do the necessary calculation when receiving input from the wireless controller.
Self-balancing robots are a common hacker project, but we don’t often see them using spinning gyroscopes to achieve that balance. Robot master [James Bruton] decided to build a robotic platform with active gyroscopic stabilization, starting from a simple proof of concept.
A gyroscope can balance, but cannot actively counteract external forces directly. However, if the gyroscope is tilted around an axis it will exert a force perpendicular to that axis of tilt, known as gyroscopic precession. By tilting the gyroscope with an actuator, and orienting the gyroscope correctly, gyroscopic precession can be used for stabilization. This is known as a control moment gyroscope. [James] demonstrated this with a 3D printed proof of concept, which is used as an IMU to measure the angle of tilt, and use a PID loop to correct the imbalance with a servo actuating the gyroscope.
His second platform used a pair of gyroscopes spinning in opposite directions to compensate for any unintended gyroscopic precession along another axis. A pair of roller skate wheels allow the entire platform to roll along. Due to a slight imbalance in the platform, [James] noticed that the gyroscopes will continue to creep in one direction, until reaching the end-stops and falling over. By adding a second software controller to keep track of how much the gyroscopes have to move to maintain balance, it can continuously calculate and update the balancing point. This prevents the gyroscopes from hitting the end stops.
Control moment gyroscopes are commonly used for attitude control on spacecraft, and to reduce the rolling motion of boats in waves. [James] has plans to combine a control moment gyroscope with the more conventional balancing method, to balance a robot on a single wheel.
When automating almost any moderately complex mechanical task, the actuators and drive electronics can get expensive quickly. Rather than using an actuator for every motion, mechanical multiplexing might be an option. [James Bruton] has considered using it in some of his many robotics projects, so he built a prioritizing mechanical multiplexer to demonstrate the concept.
The basic idea is to have a single actuator and dynamically switch between different outputs. For his demonstration, [James] used a motor mounted on a moving platform actuated by a lead screw that can engage a number of different output gears. Each output turns a dial, and the goal is to match the position of the dial to the position of a potentiometer. The “prioritizing” part comes in where a number of outputs need to be adjusted, and the system must choose which to do first. This quickly turns into a task scheduling problem, since there are a number of factors that can be used to determine the priority. See the video after the break to see different algorithms in action.
Instead of moving the actuator, all the outputs can connect to a single main shaft via clutches as required. Possible use cases for mechanical multiplexers include dispensing machines and production line automation. Apparently, the Armatron robotic arm sold by Radioshack in the ’80s used a similar system, controlling all its functions with a single motor.
Soft robotic grippers have some interesting use cases, but the industrial options are not cheap. [James Bruton] was fascinated by the $4000 “bean bag” gripper from Empire Robotics, so he decided to build his own.
The gripper is just a flexible rubber membrane filled with small beads. When it is pushed over a object and the air is sucked out, it holds all the beads together, molded to the shape of the object. For his version [James] used a soft rubber ball filled with BBs. To create a vacuum, he connected a large 200cc syringe to the ball via a hose, and actuated it with a high torque servo.
It worked well for small, light objects but failed on heavier, smooth objects with no edges to grip onto. This could possibly be improved if the size and weight of the beads/BBs are reduced.
[James Bruton]’s impressive portfolio of robots has always used conventional rigid components, so he decided to take a bit of a detour and try his hand at a soft robot. Using a couple of few inflatable pool noodles for quick prototyping, his experiments quickly showed some of the strengths and weaknesses of soft robots.
Most of the soft robots we see require an external air source to inflate cells in the robot and make the limbs actuate. Taking inspiration from a recent Stanford research project, [James] decided to take an alternative approach, using partially inflated tubes and squeezing them in one section to make the other sections more rigid. He bought a couple of cheap pool noodles and experimented with different methods of turning them into actuators. The approach he settled on was a pair of noodles tied together side by side, and then folded in half by an elastic cord. As one end is squeezed by a servo bellows, the internal pressure overcomes the tension from the elastic cord, and the “elbow” straightens out.
[James] tested various arrangements of these limbs to build a working hexapod robot but to no avail. The simple actuating mechanism was simply too heavy, and could just lift itself slightly. This highlighted a common theme in almost all the soft pneumatic robots we’ve seen: they carry very little weight and are always tethered to an external air supply. The combination of stretchy materials and relatively low pressure compressed air can only handle small loads, at least in Earth gravity and above water. Continue reading “Pool Noodle Robot Shines A Light On The Pros And Cons Of Soft Robots”→
Master of 3D printed robots, [James Bruton], plans to do some autonomous rover projects in the future, but first, he needed a modular rover platform. Everything is cooler with tank tracks, so he built a rover with flexible interlocking track sections.
The track sections are printed with flexible Ninjaflex filament. Each section has a tab designed to slot through two neighboring pieces. The ends of the tabs stick through on the inside of the track fit into slots on the drive wheel like gear teeth. This prevents the track from slipping under load. The Ninjaflex is almost too flexible, allowing the tracks to stretch and almost climb off the wheels, so [James] plans to experiment with some other materials in the future. The chassis consists of two 2020 T-slot extrusions, which allows convenient mounting of the wheel bogies and other components.
For initial driving tests [James] fitted two completely overpowered 1500 W brushless motors that he had on hand, which he plans to replace with smaller geared DC motors at a later stage.
A standard RC system is used for control, but it does not offer a simple way to control a skid steer vehicle. To solve this, [James] added an Arduino between the RC receiver and the motor ESC. It converts the PWM throttle and turn signal from the transmitter, and combines is into differential PWM outputs for the two ESCs.