Goodlancer Robotics

11/01/2025

We've successfully developed a way to control the NAO humanoid robot from an external computer—seamlessly managing both the real robot and its Webots simulation. This breakthrough means we can leverage the full power of deep learning, control theory, and advanced computational methods, bypassing the limitations of the robot's internal processor. Now, all 25 motors and three cameras are under precise control via Wi-Fi!

21/12/2024

📣We're back and we're looking for passionate volunteer team members! After a five-year break, our team is returning to RoboSport and plans to compete in the largest robot racing and sumo tournaments across Europe and Asia - including and . For the first time, we will be representing the State of Israel at these events, becoming the very first Israeli team in their history!
🤖Our goal is to participate in several European and Asian championships in 2025. If you’d like to join or help, please contact Viacheslav Nefedov Tatiana Ivanova-Nefedova or any other member of our team.
🤝We are excited to welcome new members, and we especially need a circuit design specialist right now. We’re also ready to start a club at any Israeli university willing to support us, and looking for partners and sponsors who can help us move forward. Feel free to leave a comment, and please help spread the word by sharing this post!
🧐You can find the list of topics we’re exploring - such as using AI in high-speed robotics - in the first comment under this post.

06/12/2024

Waiting for results https://robot-sumo.jp/competition-00ac692d-32df-4199-bc13-6397712d5c2c

全日本ロボット相撲大会は、ロボット力士が白熱した対戦を繰り広げる、国内最大級のロボット競技大会。ロボティクスを学ぶ最高の教材として世界30か国以上の大学や教育機関などで高く評価され、全世界で延べ8万名の方...

16/11/2024

We lowered the center of gravity of the Maqueen robot significantly to improve its balance, increase its top speed, and make adjustments easier. The standard battery holder mount was replaced with a 3D-printed mount weighing only 4 grams. It takes just 10 minutes to print and another 10 minutes to install.

30/10/2024

We have conducted an experiment to test the feasibility of teaching control theory (filters and controllers) on the Maqueen educational robot. Control theory is widely used in industry to stabilize robots, manage chemical processes, and more. We chose Maqueen as an affordable and accessible robot to explore whether complex algorithms can be run on "any robot," thereby assessing how sophisticated the hardware needs to be for practical control theory learning. Despite the robot being equipped with only three sensors, we successfully implemented the following algorithms:
- Exponential smoothing,
- Inertia compensation,
- PID control,
- Quadratic error scaling,
- Exponential smoothing,
- Complementary filter,
- Speed controller.

This represents the "core set" of control methods, with only the Kalman filter and adaptive control missing.

In the future, we aim to transfer these algorithms to a more powerful robot that we plan to custom-build for international line-following competitions. At this stage, it also seems essential to port the model with algorithms to a simulator, allowing us to fine-tune the entire set of algorithms simultaneously. We are considering MuJoCo as a simulator option, though we welcome suggestions in the comments for other possible alternatives.

24/10/2024

In the video above, you'll see our first experiment with the educational robot Maqueen by DFRobot, featuring a discrete controller with motor inertia compensation in action. We are excited to demonstrate how professional-level algorithms can be adapted to run on an affordable, accessible platform.
We are broadening our repository of algorithms for competitive robotics, and you can check it out here: https://goodlancer.com/docs-category/line-follower/

To kick off, we’ve started experimenting with the educational robot Maqueen by DFRobot, using the micro:bit controller. Our goal is to determine which of the professional automatic control algorithms can be adapted to run on an affordable, toy-like robot that's accessible even for elementary school students.

We plan to test various filters (such as exponential smoothing, complementary filters, and Kalman filters), as well as controllers (including different PID controllers and adaptive controllers). We will also explore delay-compensation models and potentially even self-learning capabilities—like parameter tuning using gradient descent and reinforcement learning.

If you're interested in this topic, be sure to follow our page and join our group on Facebook for more updates and discussions!
We’d love to hear from you! If you have interesting algorithms you'd like us to test or describe, or if you want to share your own experiments (including descriptions and videos), please drop a comment below.

https://youtu.be/w-AeYbjJlJQ