Spring-bot: A robot that uses wheels and the passive compliance of springs to vertically climb rainforest trees

Tags: Robotics

Student: Sacha Hussey
Course: Design Engineering MEng
Supervisor: Thrishantha Nanayakkara
Theme: Masters Project
Links: Sacha Hussey Email

Two percent of the world’s rainforests are destroyed annually. Non-Timber Forest Resources (NTFR), e.g. mushrooms and nuts, can be found in the upper canopy of rainforests and have high economic and pharmaceutical value. Commercialisation of these resources can incentivise governments to protect the rainforests that they are situated in. However, accessing NTFR is difficult, given their remote locations. Therefore, a tree-climbing robot is needed. This research project introduces Spring-bot, a robot that utilises a gripper, wheels, and 18 springs to vertically climb cylindrical objects. The proposed robot weighs 645 grams and can travel at a maximum speed of 6.61 cm/s.

Final Robot Prototype

Spring-bot uses a novel compliant wheeled gripper mechanism and a tail to vertically roll up trees. It utilises two arms, each with a three-link arrangement. The joints between the links are spring-loaded, so that this compliance passively takes care of force and form closure. The 18 extension springs also enable static gripping without power consumption. Actuated wheels are situated on the middle links, and passive wheels on the distal links and tail. The tail is spring-loaded to aid adhesion. Spring-bot weighs 645 grams and can travel at a maximum speed of 6.61 cm/s.

A secondary feature of the robot that was developed separately to the climbing function was the antagonistic tendon mechanism. This includes four tendons; two on each arm, creating the ability to both open and close the robot's arms.

Spring-bot can be quickly assembled using a drill, Allen key, screwdriver, and glue. With this, the components can be easily purchased and fabricated in most countries. These features were developed using a situated design approach that accounted for both the natural and socio-economic environments of the robot's context of use.

Final working robot prototype with labelled parts

Experimental Results

To evaluate Spring-bot's capabilities, eight experiments were carried out, which corresponded to the eight measures of success criteria. For climbing distance, the robot succeeded in climbing one metre on a real tree covered in sandpaper. When evaluating climbing speed, the robot achieved the measure of success of 4 cm/s, and obtained a maximum speed of 6.61 cm/s. This was tested on a 20 cm diameter blue foam cylinder. The robot's instantaneous power during climbing was measured, using two multimeters, to be 1.33 watts, passing the maximum 4 watts criteria. The robot successfully rolled over obstacles, but could not turn to manoeuvre around them, revealing that the mechanism depends heavily on wheel alignment and symmetry. To possibly obtain this capability in the future, another drive train would need to be added to enable orthogonal movement. When testing climbing diameter, the robot successfully climbed 15 cm and 20 cm diameters; accomplishing a 5 cm range, but not the 10 cm range measure of success, since it failed climbing a 25 cm diameter tree due to trunk irregularities and friction between the servo motors and trunk. The robot achieved climbing slopes from 45° to 180°, being on the underside and overside. The desired minimum payload capacity of 1 kg was not possible, however the capability of climbing with a 20 g payload was confirmed. Topology optimisation could be carried out, in the future, to increase this capacity.

Potential Future Application and its Implications

Spring-bot could be used to search for and sample non-timber forest resources in the upper canopy of rainforests, while tethered to a manual controller on the rainforest ground that is being operated by a local of the rainforest community. This application would enable the increase in NTFR collection efficiency, and therefore potentially offer a commercial incentive for governments to reduce deforestation and conserve rainforest ecosystems for the sake of increased availability of these valuable NTFR. Other implications of this application could include the creation of jobs for rainforest locals, eliminating the need for locals to risk their lives climbing trees to sample NTFR, as well as enabling governments to enforce sustainable NTFR collection through monitoring the use of the robots.

Project Details

Final Robot Prototype

Experimental Results

Potential Future Application and its Implications