💻 computer science

Durability-Aware Multi-Objective Optimization of the Jansen Linkage: Trading Gait Quality Against Joint Wear

This paper introduces a durability-aware multi-objective optimization framework for the Jansen linkage that couples kinematic, dynamic, and Archard wear models to demonstrate how adjusting link lengths can simultaneously improve gait quality and significantly reduce joint wear, outperforming the original design while remaining robust to manufacturing tolerances.

Original authors: Jichao Wang

Published 2026-06-23

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Jichao Wang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a mechanical walking leg, like the ones Theo Jansen built for his giant "Strandbeest" art installations. These legs are made of rigid sticks connected by pins (joints), powered by a single spinning crank. For decades, engineers have treated these legs like a puzzle where the only goal is to make the foot walk in a perfect, straight line. They found a set of "holy numbers" (specific stick lengths) that make the foot move beautifully.

But this paper asks a question the original designers never considered: What about the wear and tear?

Think of the "holy numbers" as a recipe for a delicious cake that tastes perfect but falls apart after one bite because the ingredients are too brittle. This paper tries to bake a new cake that tastes just as good but is sturdy enough to last a long time.

Here is the simple breakdown of what the researchers did and found:

1. The Problem: The "Perfect" Walk vs. The "Tired" Joints

The original Jansen leg is a masterpiece of motion. It walks smoothly, but the researchers realized that while the foot looks great, the pins holding the sticks together are being abused.

The Analogy: Imagine a door hinge. If you open and close it gently, it lasts forever. If you slam it hard every time, it gets loose and squeaky. The original Jansen leg design "slams" its joints with heavy loads at specific moments, even though the walking motion looks smooth.
The Missing Piece: The original design only cared about the path of the foot. It ignored the pain of the joints.

2. The Solution: A New "Fitness Test"

The researchers built a computer simulation that acts like a dual-purpose fitness test for the leg:

Test A (Gait Quality): Does the foot walk in a straight line? Is it smooth?
Test B (Durability): How much "rubbing" and "grinding" happens at every single pin?

They used a mathematical rule (called the Archard law) to calculate how much material would be shaved off the pins over time due to friction and pressure. They then used a smart computer algorithm (NSGA-II) to try millions of different stick lengths to find a version that passes both tests.

3. The Big Discovery: The "Holy Numbers" Were Actually Flawed

The most surprising result is that the famous "holy numbers" (the original stick lengths) are not the best they can be.

The Metaphor: It's like finding out that the "perfect" recipe for a car engine actually wastes fuel and wears out pistons faster than necessary.
The Result: By tweaking the stick lengths by a small amount (less than 30% change, so it still looks like the same leg), they found a new design that:
- Walks better: The foot is flatter and moves more smoothly.
- Lasts longer: The total wear on all the pins dropped by 56%.

This means the new leg could theoretically last 2.3 times longer than the original before the joints wear out, all while walking better.

4. How They Did It (The "Magic" Behind the Curtain)

To prove this wasn't just a lucky guess, they did three things:

Cross-Checking: They ran the math two different ways (like checking your work with two different calculators) to make sure the forces were calculated correctly.
Stress Testing: They tested the new design at different speeds and with different weights (payloads). The new design remained superior in almost every scenario.
The "Sensitivity" Check: They figured out which specific sticks were the "bosses" of the wear. They found that just two or three specific stick lengths controlled most of the friction. Changing these was the key to the improvement.

5. The Catch (Real-World Reality)

The paper is honest about its limitations:

Ideal vs. Real: The computer model assumes the pins fit perfectly with no gaps. In the real world, metal parts have tiny gaps and manufacturing errors.
The Tolerance Test: They simulated what happens if the sticks are cut slightly wrong (like a carpenter making a tiny mistake). The new design is still better, but if the manufacturing isn't precise, the "walking quality" might suffer. The new design is more sensitive to being built perfectly than the old one.

Summary

This paper is a "tune-up" for a famous mechanical leg. It proves that by slightly adjusting the length of the sticks, you can make the leg walk smoother and last twice as long. It shows that the original "perfect" design was actually missing a crucial piece of the puzzle: durability.

Key Takeaway: You don't have to choose between a leg that walks well and a leg that lasts long. With the right math, you can have both.

Technical Summary: Durability-Aware Multi-Objective Optimization of the Jansen Linkage

Problem Statement
The Jansen linkage, a single-degree-of-freedom planar leg mechanism famous for its use in Theo Jansen's Strandbeest, has historically been designed solely to optimize foot-path gait characteristics (flatness, ground clearance, absence of cusps). The original "holy numbers" (link lengths) were derived via evolutionary search based exclusively on kinematic criteria, with no consideration for the durability of the revolute joints. In physical implementations, joint wear—governed by the product of contact load and sliding distance—ultimately limits service life. Previous research has focused on kinematics, dynamics, and gait optimization, or on wear prediction in generic mechanisms with clearance joints, but joint wear has never been adopted as an explicit design objective for the Jansen leg. This paper addresses the gap by introducing a durability objective into the dimensional synthesis of the linkage.

Methodology
The authors developed a fully parametric pipeline coupling three distinct models to evaluate any set of link lengths:

Kinematic Model: A forward-kinematic solver using the two-circle-intersection method to determine joint positions and the foot trajectory based on the crank angle.
Dynamic Model: An inverse-dynamics formulation using a seven-body, ten-joint system. Joint reaction forces are calculated via a constraint-Jacobian and Lagrange-multiplier approach (Newton-Euler equations). The model treats joints as ideal, clearance-free holonomic constraints and applies a prescribed vertical ground-load during the stance phase. This formulation was independently cross-verified using a reduced single-degree-of-freedom energy method.
Wear Model: An Archard wear model estimates the volume of material removed per gait cycle at each pin. The calculation uses the cycle-mean reaction force, the pin radius, and the per-cycle relative rotation. The study distinguishes between "slip-dominated" pins (crank bearings with full 360° rotation but low load) and "load-dominated" pins (grounded bearings with high load but small oscillation).

These models were integrated into a bi-objective optimization problem solved via the NSGA-II algorithm. The objectives were to minimize:

$f_1$ : A composite gait error (weighted sum of stance flatness and stance horizontal-velocity ripple).
$f_2$ : Total joint wear (sum of wear at all ten pins).

The optimization was subject to constraints on step length, ground clearance, duty factor, and assembly feasibility, with link lengths bounded to $\pm30\%$ of the original Jansen values.

Key Contributions

Parametric Evaluator: The creation of a coupled kinematic-dynamic-wear evaluator for the Jansen leg, validated by an independent energy-based dynamic formulation.
Pareto Dominance Discovery: The demonstration that the classical Jansen "holy numbers" are Pareto-dominated when durability is included as a design criterion.
Robustness Analysis: A sensitivity study identifying specific link lengths that govern wear variance and confirming the robustness of the optimization results across a wide envelope of crank speeds and payloads.

Results

Pareto Front: The optimization produced a Pareto front where every non-dominated design simultaneously improves gait quality and reduces total joint wear compared to the original Jansen design. The classical design is dominated by all solutions on the front.
Performance Gains: A representative optimized design (with link length adjustments within $\pm29\%$ $\pm 29%$ ) achieved:
- A 28% improvement in stance flatness.
- A 58% reduction in stance velocity ripple.
- A 56% reduction in total joint wear.
- A 59% reduction in peak joint wear.
- A 19% increase in step length, at the cost of a 15% reduction in ground clearance (which was an active constraint).
Wear Distribution: The reduction in wear was broad-based, occurring at every pin (ranging from -20% to -85%). The most significant reductions occurred in load-dominated pins. The remaining wear is largely dictated by slip-dominated crank bearings, which form an irreducible "wear floor" independent of link geometry.
Sensitivity and Robustness:
- Operating Conditions: The wear reduction (48%–56%) remained robust across a crank-speed $\times$ payload envelope, except where inertial forces dominated at high speeds and low loads.
- Link Lengths: A Sobol variance-based analysis identified link lengths $c$ and $k$ as the dominant factors governing total wear variance. Lengthening $c$ and $e$ was found to most effectively reduce wear.
- Manufacturing Tolerances: A Monte-Carlo study showed that the wear advantage degrades gracefully under realistic fabrication errors (retaining a ~43% reduction at $\pm1\%$ tolerance), though the feasibility of the design is sensitive to clearance constraints.

Significance and Claims
The paper claims to provide a practical route to longer-lived walking linkages by demonstrating that durability can be integrated into the dimensional synthesis of single-DOF mechanisms without sacrificing gait quality. The authors assert that for the Jansen linkage, gait quality and durability are not conflicting objectives in the explored design space; a single redesign secures both.

The study explicitly limits its claims regarding absolute life prediction. Because the wear model assumes ideal, clearance-free joints, the resulting figures represent a relative comparative ranking rather than an absolute life prediction. The authors note that the 56% reduction in per-cycle wear projects to an idealized relative upper bound of roughly 2.3 times the joint service life under unchanged material assumptions.

The framework is presented as mechanism-agnostic, applicable to other single-DOF leg linkages (e.g., Klann mechanisms or eight-bar designs). The authors identify limitations including the idealized joint model, the mean-force approximation in the wear calculation, and the lack of experimental validation, noting that future work will focus on coupling wear-induced clearance growth with impact dynamics.