Report & Slides

[EugenioBugli]

TODOS:

• Introduction
• Mpc formulation
• Obstacle generations
• Simulation & Results
• Possible Improvements and conclusion

[salvatore373]

Report Detailed Schedule (Chapters and SubSections):

Each section contains additional information so that, even if it has multiple contributors, the style is unitary throughout the report.
All the formula must have punctuation at the end.

• Introduction
  • Task Description
  • ?Related works?
  • Solution proposed by this paper and how we dealt with it
• 3D-LIP Model with Heading Angle
  • Overall description
  • Local and Global RFs: how and why state and input are transformed between local and global RFs
• LIP-MPC: Gait planning with Model Predictive Control
  • Intro: why, what it tries to achieve
  • Heading Angle preprocessing: Describe why theta/omega are preprocessed and how it is achieved in our project
  • Kinematic Constraints (for each constraint: meaning and expression).
    • Intro
    • Walking Velocity
    • Leg Reachability
    • Maneuverability
  • Control Barrier Functions
    • General Definition, what used for
    • LCBF: why introduced, how defined in the paper
      • Our variant: what does it mean, why useful, what does it achieve (HumanoidMPCCustomLCBF, commit bbf26d0)
  • Proof?????
• ?Integrations to the paper?
  • RRT/RRT* for subgoals
  • LiDAR for unknown environment
• Simulations
  (for each simulation:
  -- plots of evolution of state, input, position error, ?optimization function?;
  -- simulation as grid of TBD frames of size TBD or simulation as single figure containing the whole path and all the footsteps;
  -- plots as PDF; keep always the same colors/style; each figure must have a description;
  -- explain the goal of the simulation, the meaning of each plot, the achieved results, which components of the project are involved in that solution)
  • Simulation without obstacles;
    goals: on same x axis on right of Xo, on same y axis on top of Xo, at position (w, w) w>0
  • Simulation with multiple random obstacles;
    goals: on same x axis on left of Xo, on same y axis on bottom of Xo, at position (w, w) w<0; (if the robot get too close to obstacles, talk about our CBF)
  • Simulation with narrow passage: MPC finds solution with their CBF but not feasible -> solved by our variant
  • Simulation with obstacle between Xo and the goal -> no solution with MPC -> comparion with solution using RRT and subgoals
• Conclusion
  • briefly, goal of our project, content of the paper, what we realized, what results we obtained, how our integrations improved results.

[EugenioBugli]

Base simulation:

• lot of obstacles from south west to north east
• different timestamps of a simulation

Edge cases:

• starting point surrounded by obstacles
• ending point surrounded by obstacles
• starting/ending close to an obstacle
• wall (RRT)
• maze with obstacles