We have exciting new projects, not only for SA and MA students, but also for freelancers looking for a fun and interesting project next to ETH obligations. If you are interested in a SA or MA please contact one of our PhDs to discuss project ideas. For Freelancer projects, check out the list of current projects below and contact one of our TAs for more information.

If you are interested in previous projects, go to our Project Report site, where you can download multiple reports of previous projects.
If you are from a different team and your team works with the B-Human code framework, make sure to check our YouTube channel for some tutorials on the framework by Filippo Martinoni.

Fall 2019 Project Proposals

Realtime Object Detection on Embedded Platforms [taken]
Integration of a Walking Engine on a bipedal robot for RoboCup [taken]
Real time ball motion estimation in RoboCup [open]

Current Freelancer Projects

This is a list of most current freelancer projects. If you are interested or have other ideas, please contact one of the team’s TAs for more information.

  1. Improve Whistle Detection [open]
    At the beginning of the game or after a goal is scored, all robots listen for a whistle and start playing when they hear it. If they don’t hear it, they have to wait for 15 seconds before they can start playing. Unfortunately, our current whistle detection is not up to par. Sometimes we don’t hear the whistle at all and sometimes we hear false whistles, which also results in a penalty. We would like a new (and better) whistle module for the upcoming tournaments!
  2. Improve the Goalkeeper Behavior [open]
    The goalkeeper’s ultimate job is to prevent the ball from being anywhere close to the goal; look at whether the current behavior does that well. What happens when the ball is not in sight, when the ball is seen far away, when the ball is near, after blocking the ball, etc. How does it react to different types of attack from opponents (center shot, goal corner shot, weak shot, saved shot that bounces back, …)?
  3. Faster Approach and Kicking of the Ball [open]
    Our robots currently take considerable some time to walk to the ball, align to the right direction and kick it in the right direction. Find out which parts of the ball approach are slow and make them faster.
  4. Global Positioning [open]
    Having good robot detection, knowledge about the location of our own players and a shared ball location, the positioning during searching and general path-planning can be optimized using this information as well. The goal is to create a strategy for each individual robot that would allow to 1. optimize their positioning on the field when searching and 2. avoiding walking into other robots. 
  5. New Kick Development [open]
    During the 5 vs. 5 matches of our NAO robots it often happens that our robots need to duel with an opponent for acquiring the ball. In these cases, a side kick can be beneficial to use if the ball is stuck between our robot and the opponent, e.g. when dribbling towards the opponent. Therefore, we would like to create a new side kick motion, which can be executed in these situations. The motion of the kick should first be developed in a motion simulator. 
  6. Better Flipping Detection [open]
    As we localize on the field only using the field lines that are symmetrical, it is possible for the robot to flip on the wrong side and still maintain a high confidence, e.g. after falling. We would like to find a more reliable way to detect flipping through an intelligent comparison and merging of the own perceptions with these of other teammates.
  7. Improved Parallelization [open]
    Previously computational power on our Nao platform has been limited to a single processor thread, but the newest version (V6) of our robots has multiple cores and integrated processor graphics. To enable new projects on our limited and real-time embedded platform, mostly in the area of machine learning and computer vision, we want to improve performance by better usage of parallelization across our entire framework.
  8. Balanced Standing [open]
    When playing for longer times, the joint motors of the robots are subject to overheating and imparing the robot’s walking. Therefore, we want to shut off the robot’s motors when they are far away from the ball and waiting – e.g. for the goalie and defending players. For this, we need to find an equilibrium where the robot is able to stand without the need to continuously control the joints. However, it still has to be ensured that the motors switch back on when the robot is imbalanced or needs to move.

Spring 2018 Project Proposals

  1. An Advanced Software Framework for Real-time Multi-Robot SLAM
  2. Coordinated Defensive Gameplay for RoboCup

Fall 2017 Project Proposals

For the fall semester all projects have already been taken. If you are interested in a project for winter/spring 2018, please don’t hesitate to contact one of our TAs.

  1. Detecting own and opposing robots
  2. Game play strategy and position optimization
  3. AI / Robot Communication / Computer Vision
  4. Dynamic walking on different surfaces
  5. Improvement of robot localization during game play (i.e. the symmetry problem)

Spring 2017 Project Proposals

  1. Robot external calibration
  2. Top View Detection and Tracking of Field Elements
  3. Robot Communication between Coach and Players
  4. Improving Self-Localization
  5. Improvement in Kicking and Passing Behavior
  6. Adding a Model Predictive Controller to the Passing Engine

Fall 2016 Project Proposals

  1. SLAM continuation
  2. Self-localization
  3. Ball detection
  4. Decision making for behavior (e.g. shoot, dribble, pass)
  5. Collaborative ball localization
  6. Automated color calibration
  7. Gaze stabilization for NAO robot

Spring 2016 Project Proposals

  1. SLAM continuation
  2. Collaborative ball localization
  3. White goal detection and localization
  4. Robot classification
  5. Robot detection
  6. Decision making for behavior (e.g., shoot, dribble, pass)
  7. Gaze stabilization for NAO robot
  8. Master project: Dynamic Analysis and Optimal Control of a Robotic Bipedal Locomotion System