Terrain Conforming Trekking Pole

Roughly 764,000 people in the United States live with cerebral palsy, about 500,000 of them under eighteen. The condition affects gait in several distinct patterns, true equinus, jump gait, apparent equinus, crouch gait, and each one changes how a person loads and balances on uneven ground.

We started the year by reading what people actually say about hiking with CP, across disability forums, trail communities, and advocacy writing. The recurring theme was not fitness. It was attention. One hiker described never being able to look up at the trees because of constantly watching the ground to avoid tripping, calling it mentally taxing. Another wrote about being left out because a hike was assumed to be too hard for them.

Our interview with a fellow student who has CP was filled up with the same frictions repeated in different words: stairs, rocky trails, having to concentrate hard not to trip on long walks, years of physical therapy, and a strong preference for equipment that does not announce itself. That last one mattered more than we expected, discretion is a design requirement, not a nice to have.

Prior art fell into two camps. Heavy adaptive mobility equipment solves stability but is conspicuous and trail limited. Standard trekking poles are light and socially invisible but do not adapt: the tip is rigid and normal to the shaft, so any angle between it and the ground turns downward force into a sideways slide. We chose to keep the familiar form factor and change only what happens where it meets the ground.

Granular jamming is the effect behind a vacuum-packed bag of coffee. Loose grains flow freely at atmospheric pressure, but pull the air out and they lock into a rigid solid that holds whatever shape it was in. Applied to a pole tip, that gives exactly the two states we needed from one part.

The sequence is one gait cycle. The pole comes down with the membrane soft, so it drapes over the root or rock it lands on. Sensing switches in the coupler skirt detect contact and fire the vacuum pump, which evacuates the membrane and locks the conformed shape under load. The user pushes off, the pole lifts, and the membrane returns to flexible.

Traditional poles get a point. Ours gets a footprint.

Before settling on a pole we sketched the obvious alternatives. A knee brace with a coil spring and a nylon pull cable, where moving the base spring point adjusts resistance. A passive exoskeleton with a thigh support, torsion springs, and an anchored knee joint. And a shoe with a rack and damper mechanism in the sole, using a one way rotary damper to absorb the impact of each step.

All three had the same problem: they are worn, which means they are conspicuous, they take time to put on, and they change how the body moves rather than what the ground feels like. Users had told us plainly that they did not want equipment that announced a disability. A trekking pole is already accepted hiking gear, so we chose to hide the technology inside something people were happy to be seen carrying.

The most mechanically involved of the three put the damping in the sole: a rack and retainer driving a one way rotary damper, sandwiched between a spongy upper and a hard sole, so each step is absorbed on the way down but returns nothing on the way up. It was the clearest illustration of why we went the other way, a shoe has to be manufactured to fit, replaces something the user already owns, and puts a mechanism under full body weight thousands of times a day.

The first drawing already had the shape of the final product: a switch at the grip, an air tank and actuator on the shaft, tubing down to a granular gripper, and the two states the tip moves between as it meets the ground.

The first working proof of concept was deliberately crude: an ice pack membrane filled with coarse coffee grounds, air tubing, and a manual trigger. It jammed. That was enough to justify the whole approach and put us in front of users with something they could hold.

We took it to Hod Lipson, Columbia professor and one of the pioneers of granular jamming, for feedback and got three lines that shaped the rest of the year: easy to do, hard to do well, start simple because it's gonna get messy. All three turned out to be accurate.

Early user feedback was encouraging on the fundamentals, lightweight, auto-actuated, height adjustable, and practical for balance on rough terrain, and specific about what to fix. People with limited mobility walk slowly, so actuation timing mattered less than we had assumed. Buttons needed to be light enough to press easily but not so light they trigger by accident. Keeping total weight down was the highest priority, and the tip surface area could come down slightly.

The membrane is the part the whole product lives or dies on. It has to be thin enough to conform, tough enough to survive thorns and rock, and elastic enough to recover thousands of times. We tested balloon, condom's latex material, and gel ice pack membranes off the shelf, then started casting our own.

Using spin casting jigs we produced latex, urethane, and silicone membranes across two durometers (30A and 60A) and compared them on conformity, recovery, and puncture resistance. The eventual pick was a commercial stress ball membrane in thermoplastic rubber with a four inch unstretched diameter, filled to a 95% packing fraction with coffee grounds, which held up better than anything we cast.

Everything between the pole and the membrane is a custom 3D printed coupler, and it has to route air, route wires, filter grounds out of the pneumatic line, and carry the user's full load into the shaft. It resolved into three parts: a core with an air channel and a wire channel, a retention collar, and a skirt that houses the sensing switches.

The air path runs 3/8 inch NPT push-to-connect fittings and nylon piping from the pump down through the core, with a replaceable filter slot at the membrane end so grounds cannot reach the pump. The whole assembly bolts together with 6-32 screws into heat-set inserts, which was a deliberate modularity call, every subsystem had to come apart for repair without cutting anything.

The FEA is only as good as the material model, and a 3D printed part is not the PLA on a datasheet. Infill pattern, wall count, and print orientation all change how it behaves. So we printed tensile specimens to our exact settings, four solid walls at 70 percent gyroid infill, and pulled them to failure on an Instron Universal Testing Machine.

Stress-strain analysis of that data gave us the numbers the simulation needed. A linear elastic fit over the elastic region returned a Young's modulus of 2782.6 MPa, and a 0.2 percent offset construction put yield at 24.40 MPa. Plotting transverse against axial strain over the same region gave a Poisson's ratio of 0.309. Those are the properties of our material at our print settings, not a generic PLA.

With measured properties in hand the coupler was loaded in simulation the way it is loaded on trail. Trekking poles are rated to take about 30 percent of body weight, so we sized the coupler against 266.6 N applied at 45 degrees to the worst case face. FEA using those measured properties put peak principal stress at 3.4 MPa, converged to within one percent across a mesh study.

The exploded view is the modularity doctrine made literal: pole, coupler, membrane, pump housing, electronics enclosure, fittings, and piping all separate. Nothing is glued and nothing is captive, so a torn membrane or a dead pump is a five minute swap rather than a rebuild.

The working prototype: trekking pole, electronics enclosure below the grip, pump housing clamped mid-shaft, air line running down to the coupler, and the granular membrane at the tip.

The electronics are an Arduino controller, an H-bridge driving a 12 V vacuum pump rated to 81 kPa and 10 L/min, sensing switches, and a rechargeable battery, all inside a housing clamped to the pole and built to NEC 725 low-voltage wiring and IEC 62133 battery safety standards.

Battery sizing started from volume, not current. We estimated the air volume in the tubing, the coupler core, and the membrane at a 95 percent fill fraction, divided by the pump's volumetric flow to get a time to vacuum, then combined that with published stride times (around 1.14 seconds) to work out what fraction of each step the pump actually runs. Continuous operation projected 6.5 hours on a 6000 mAh pack, duty-cycling to half the gait cycle nearly doubled it to about 12.

In practice the pump drew considerably less current than the manufacturer's rating suggested, so the real figure sits at the optimistic end of that range.

Puncture testing assumed a hiker puts about 20 percent of body weight through a pole, or roughly 40 lb, and we pressed rocks, sticks, thorns, and nails into the membrane at and above that load, then walked the assembled pole over a bed of the same debris. The membrane held.

Beyond puncture, the membrane took a 54 lb (240 N) compressive load and survived a drop test, and the printed coupler sustained a 50 lb (222 N) shear load without failure, consistent with the FEA. The final prototype demonstrated granular jamming under real operating conditions.

The honest caveat is that most of this was qualitative. We proved the mechanism works. Instrumented validation, coupler optimization, and real field testing on trail are the next things anyone building on this needs to do.

We presented TerraLock at the Columbia Engineering Senior Design Expo alongside a poster and the working prototype, after six design reviews across two semesters.

What I would carry forward: the crude ice-pack prototype taught us more in a week than a month of CAD did, and the modularity doctrine, every subsystem removable with a hex key, is the only reason we could iterate the membrane six times without rebuilding the pole. The remaining wish list is reversible airflow, a smaller and lighter electronics package, a universal pole fit, and enough user facing adjustment that a hiker can tune the tip to their own gait.