The domain of expertise required to fully analyze this material is Applied Robotics and Precision Mechanical Engineering, with a strong focus on real-time control systems and motion capture technology.
I will now adopt the persona of a Senior Systems Integration Engineer specializing in High-Speed, Low-Latency Motion Control.
Abstract:
This video documents the ambitious, multi-stage engineering effort to construct an electromechanically complex, auto-aiming golf putter designed to correct a user's swing trajectory in real-time to ensure the ball enters the hole. The primary objective is to overcome the skill deficit against a proficient opponent (the creator's wife) using technology. The project encountered significant challenges across mechanical design, control theory application, and system integration, moving from initial concepts to a functional, albeit limited, prototype.
Key engineering hurdles included: achieving the necessary precision (corrections within the $\approx 200$ millisecond window of a human swing); managing the rotational kinematics by pivoting around a vertical axis without ground collision; mitigating reaction torques through the implementation of a dual-head, counter-rotating inertial balance system, which was initially flawed due to unconsidered motor inertia amplification via gearing; and establishing a high-fidelity spatial tracking system using OptiTrack cameras to determine the precise 6D pose of the club face, ball, and—later—the hole components.
The final demonstrable prototype successfully executes basic, straight-line corrections, primarily constrained by a system latency issue (exacerbated by Python/Windows implementation) that leads to significant overshoot. Furthermore, custom-built putting surfaces introduced new, unmodeled physics complexities, such as wall-bounce energy transfer and course-curvature effects, necessitating future upgrades to full physics simulation for advanced play (e.g., bank shots). The initial test run resulted in a qualified success (a 6/10 score from the judge/wife), signaling the project's transition into a "Part Two" phase focusing on software upgrades.
Auto-Aiming Robotic Putter Development: System Integration Review
0:00:02 Motivation & Goal: Develop an "auto-aim" mini golf putter to compensate for amateur swinging technique, enabling accurate ball guidance into the hole, even in darkness.
0:01:04 Precision Requirement: The system must execute tracking, calculation, mechanical aiming, and correction within the $\approx 200$ millisecond window of a standard swing.
0:01:42 Fundamental Mechanism: Control the ball trajectory by precisely modifying the angle of the club face during the swing, achieved via pivoting around a vertical axis beneath the putter head.
0:02:38 Mechanical Solution – Drive Train: A motorized system uses a drive shaft running down the club shaft, coupled with a worm gearbox at the head. The worm drive provides a one-way lock, preventing impact forces from twisting the drive shaft.
0:03:53 Inertial Torque Compensation (Initial Design): To counteract the reaction torque generated by rapidly twisting the putter head, a secondary, counter-rotating inertial mass (counterweight) was introduced to keep the main club handle stable in the user's hands.
0:04:56 Prototyping Fabrication: The putter head was manufactured via 5-axis CNC machining, while other components utilized water-jet cutting and SLS 3D printing (nylon).
0:05:56 Prototype Configuration: The initial working model is scaled down for desk testing. It features a high-torque, low-speed BLDC motor controlled by an ODrive, using the worm drive to actuate the putter head.
0:06:31 Torque Management Refinement: The initial counterweight design caused lateral acceleration/movement of the club head. The solution involved modifying the motor system by adding inertia rings to the motor itself, utilizing the motor's reaction torque to cancel the torque from the putter head actuation, thereby eliminating the need for the external inertial counterweight.
0:07:11 Design Files Availability: All design files for the desktop prototype are available on Onshape for free access.
0:07:44 Initial Test Failure: The first full actuation test showed the club head moving laterally instead of purely rotating, confirming the initial inertial balance was incorrectly applied.
0:08:43 Root Cause Identification: The motor's reaction torque, amplified 30x by the gearing ratio, was significant and had been ignored. This torque had to be actively canceled by adjusting the motor's own inertia.
0:10:38 Custom Course Construction: Modular mini golf holes were constructed from wood (MDF top for flatness), steel tubing perimeter for consistent bounce, and glued-down turf.
0:11:13 Unforeseen Physics Issue (Course): Testing revealed that the custom holes were not flat, having a $\sim 1$ cm bowl-like depression, causing the ball to deviate sideways, complicating prediction algorithms.
0:11:57 Spatial Tracking System: An OptiTrack system (multiple IR cameras and reflective markers) was implemented to precisely track the 6D pose (position and orientation) of the ball and club head at 240 Hz (0.5 mm accuracy).
0:13:20 Club Face Pose Calculation: A 3-marker jig affixed to the putter face was used to define the plane, allowing computation of the exact, offset position of the striking face relative to the tracked markers.
0:14:06 Ball Tracking Modification: The ball was covered entirely in reflective tape to act as a single large marker for reliable tracking, compensating for small dots being unreliable.
0:14:48 Initial Straight-Line Test: The system was configured for the simplest case: aiming the ball straight into the hole, ignoring complex banking or turf curvature.
0:15:18 Control Activation: A single button press initiates tracking, swing motion prediction (6D trajectory fitting), speed modeling, and real-time adjustment for user hand deviations.
0:15:54 Test Result (Initial Integration): The first successful run hit the ball in the "wrong direction," indicating "integration hell." Numerous subsystem failures (motor cogging, tracking interference, fried controller, ratio errors) were resolved after 1,000+ tests.
0:17:16 Critical Latency Issue: The primary failure mode involved timing discrepancies between the low-latency motor controller and the higher-latency external OptiTrack/Python processing pipeline, causing the system to overcorrect based on stale position data.
0:18:22 Latency Hack: Instead of rewriting the control loop in C++, a software hack was implemented to time-align the motor command age with the camera data age, synchronizing the inputs to achieve stability.
0:19:09 Final Straight-Line Success: The system demonstrates successful auto-aiming for straight shots, including aiming at a wall and having the club redirect the ball into the hole.
0:21:36 Limitations & Future Work (Part Two): The current system cannot handle complex mini-golf elements (bouncing off walls, loops), requires sufficient swing speed, and fails on very short/fast swings. Achieving true mini-golf requires a major software upgrade to a full physics simulation (including rolling friction and wall rebound dynamics), which is computationally intensive (25-hour computation time between shots reported for the initial physics model).