My name is Tyler Veness. I'm an embedded software engineer with a passion for controls engineering and teaching software/controls engineering to high schoolers in the FIRST Robotics Competition.

Projects

Most of my free time is devoted to writing software for robotic applications. I contribute to WPILib, I wrote a book on modern control theory, and I do numerical optimization R&D.

A portfolio of my projects is available on GitHub.

Sleipnir

• A linearity-exploiting sparse nonlinear constrained optimization problem solver that uses the interior-point method
• Written in C++23
• Uses custom symbolic autodiff with linearity detection to cache results
• Uses Eigen for sparse linear algebra
• Beats CasADi + Ipopt by 24.52x on quadratic problems and by 3.39x on nonlinear problems

Controls Engineering in FRC

• A practical guide to controls engineering for high school students
• Covers:
• Fundamentals of control theory
• Modern control theory (state-space, LQR)
• Kalman filters
• System modeling (Newtonian and Lagrangian mechanics)
• Motion planning (1 DOF profiles, trajectory optimization)
• The latest PDF is available here

WPILib

• A C++ and Java library used by about 3000 high school FIRST Robotics Competition teams
• Helped modernize C++/Java code base and APIs (JDK 11, C++20)
• Maintained and contributed heavily to wpimath subproject and facilitated community contributions
• Geometry (Translation2d, Rotation2d, Pose2d, Translation3d, Rotation3d, Pose3d)
• 2D spline trajectory generation and constraints
• DARE solver that's faster than SLICOT on x86-64 desktop
• 1.48x faster with dynamic matrices and precondition checks
• 4.63x faster with static matrices and no precondition checks
• LQR and model-based feedforwards
• Kalman filters (linear, extended, unscented) and pose estimators
• Numerical integration (RK4, RKDP)

wpiformat

• Maintained style guide for WPILib projects and wrote wpiformat Python module to automate conformance
• Used Google's cpplint.py, clang-format, and custom checks and transformations