“Redesign the ‘Stigmergic Ant Hexapod Robot’, which was designed during the first semester of scholastic year 2011-2012 as an eye-catcher for the Industrial Design Center (IDC) on expos. The main goals of the redesign are to use lighter and more durable materials, make the robot modular and repair-friendly, make the robot autonomous and solve the mechanical issues of the current robot, such as burning motors, gravitational separation, high energy consumption and so on. Integrate a nice interaction aspect which fits the context of expos.”
Strictly using digital production techniques (3D printing, laser cutting and thermoforming on CNC – milled molds), a modular and reproducible, scorpion-based hexapod robot was created. Problems such as gravitational separation and high energy consumption of the motors were solved by the use of elastics, based on the principle of antagonistic muscles. The tail actuation and main robot structure were strongly based on nature’s solutions as well.
The scorpion is programmed with a range of moves and interactions to function more autonomous on expos, while a GUI allows easier calibration and live control over the robot. A marker integrated in the tail allows the scorpion to “stab” expo visitors, creating a trace of persons ‘carrying the information shared at the IDC-stand’ through the expo and representing the idea of stigmergy.
Presentation Movie : ScorpionHexaPod_UGent
Team: Flamand Stephan, Terryn Robbe, Deconinck Pieterjan |Industrial Design Center | Mentor: Saldien Jelle
We had some trouble with loosing power when using the Lipo battery. This didn’t occur when using a lab power supply. After a lot of measuring, we concluded the UBEC (DC-DC converter) got triggered into over current protection. Apparently, this gets triggered faster with a Lipo battery then with a switching power supply. Therefore, a second UBEC was added to divide the power. The problem was fixed after adding this second UBEC.
In the topshell, 4 sensor were added for interaction with the public. 1 long range IR sensor is used to detect nearby people in front of the scorpion. 3 close range sensors are used to react to a hand going over the back of the scorpion. With a state machine, the behaviour of the scorpion is determined based on the sensor inputs.
So, there are 2 main codes. Full code with serial communication to pc. This is to be used with the GUI. The GUI provides many possibilities: calibration, walking/turning, body moves, extra behaviour (claws, tail) etc.. With this, new moves are easily programmed. The second code is the autonomous part, based on the sensors.
We also added new movement possibilities. Walking can now be done in all 360° directions (eg. forward, backward, 90° left, 45° right, 135° left back…). Additionally, the leg movement is made relative to the body, so walking and turning can be done in every body position.
When testing the hexapod robot, we noticed that the movements are much smoother on a floor with a lot of friction… The robot will always be used on expos, etc. That’s why we cut a carpet of 1.5m diametre for the scorpion robot. In this way, he can be put on the ground and the friction with the ground will make the movements even smoother and more spectacular…
A lot has changed since the last update. In the following video we give an overview of the updates on the robot/software:
In this video we demonstrate the effect of the elastic strings on the energy consumption. In comparison to the last video (3 to 4 A for default position), the current draw has been lowered drastically (0 to 1 A in default position, depending on activity). This means battery live span will be a lot longer.
The addition of body inverse kinematics is shown. This offers 6 degrees of freedom for body moves. In combination with the leg inverse kinematics, this offers a big range of motion options.
Notice the tests in the video are performed on a small carpet. This is done to give extra grip to the legs. Without, the energy consumption is higher, because the legs slip a bit. This could be solved by adding sticky ends (rubber,…) to the legs.
Since the body of the scorpion is close to being finished, we can now really start programming all the moves on the robot itself. This is a first test of the walking (half speed). This will be tuned in the next days. Other moves and tricks are in development.
A GUI has been made to make programming and debugging a lot easier and faster. With this GUI calibration values and standard moves can be performed very easy and fast. For demos, the scorpion will be programmed to function autonomously.