Search

Legged robots have been proposed as a possible alternative to wheeled systems for accessing so-far hard to reach locations in our solar system. However, despite promising results in this field, significant challenges have to be solved before deployment. In particular, the actuation of walking robots in space is still an open engineering question. In this work, we focus on actuation of space-legged robots and propose a preliminary actuator sizing. Torque and velocity profiles are derived from simulations, and power consumption, heat generation and emission are computed as functions of several design parameters. A simplified thermal model is used to calculate the equilibrium temperature. Finally, design specifications and considerations are drawn from the analysis.

Identifying robotic traverses on the surface of other celestial bodies is essential to assess the capabilities of the required system at the mission planning stage. With increasingly diverse robotic systems designs for space, including wheeled, walking, and multimodal systems, a wider range of behaviors concerning electrical energy consumption and failure risk are becoming available. Thus, it becomes necessary to define path optimality for the two parameters individually, beyond a traditional minimization of path length. This paper proposes a path planning algorithm that finds optimal global paths on the lunar surface for robotic energy consumption and risk, where the user can define the energy and risk minimization functions and their relative importance. Based on a custom A* implementation, the proposed algorithm successfully minimizes the energy consumption and path risk in various scenarios. Exemplary, a cost function for the walking robot ANYmal was generated in simulation and applied to our planner. The results show that different optimized global paths were generated depending on the user’s energy/risk trade-off.

So far, mobile, robotic exploration of celestial bodies has been performed solely with wheeled systems, making access to highly unstructured, compressible, and sloped areas very challenging. Additionally, the systems have to account for unknowns, such as unclear soil properties or the environment in uncharted areas such as caves. Scientists, however, have identified those challenging places as potential targets for future missions. Meanwhile, versatile, terrestrial walking robots have advanced rapidly over the last decade and have reached a maturity level where the commercial application becomes viable. Thus, it is only a matter of time until walking robots will be used to explore so far unreachable areas in our solar system. Nevertheless, it is still unclear which underlying technologies need to be adapted and further developed for space and how such a robot would look like. In this thesis, we investigate legged systems in the context of space exploration. We were particularly interested in increasing and validating the system performance in partially unknown, sandy, sloped, and lower-than-earth gravity settings. We analyze the state-of-the-art of legged robots and the scalability of underlying technologies in view of future missions and carried out initial simulation studies and field trials. The studies revealed that dynamic walking gaits, especially those with extended flight phases, become increasingly efficient in low-gravity. Although the tests also showed that large contact forces lead to deep sinkage and ultimately increased the risk of falling. We introduce the quadruped robot SpaceBok and respective hardware modifications to perform efficient, repetitive, low-gravity jumping. To this end, we installed a reaction wheel that allows for stabilization during the extensive flight phase and parallel elasticities in the legs to increase energy efficiency. To tackle the sand-walking challenge, we developed novel passive-adaptive planar feet, which reduced sinkage and increase traction on highly compressible, dry granular media. Together with a terrain-adapting walking controller, we demonstrate the traversal of steep, representative Martian analog slopes. We evaluate the robot's performance in terms of safety and efficiency and propose effective strategies for overcoming slopes. Lastly, we present an approach to perform a tactile inspection of the terrain in front of the robot using one of the limbs. This inspection-before-traversal increases the certainty about the terrain type and thus the safety of the system. We spun off this approach to the terrestrial application of concrete inspection in sewers. Here, we performed semi-autonomous missions in a subterranean environment where a tactile inspection motion was carried out every few meters to assess the state of deterioration. The technology developments presented in this thesis have been thoroughly tested and validated on two advanced quadrupedal robots, namely ANYmal and SpaceBok.

In this workshop, we reviewed the state of the art and provide an outlook towards the future of robotic, planetary surface exploration. As the scientific community strives to perform in-situ investigations of the unexplored areas in our solar system, requirements on mobility are becoming increasingly demanding. Steep craters, caves, highlands, permanently shadowed regions, and icy worlds are only a few examples of extreme environments a potential robot has to account for. In this regard, it is essential to share and discuss insights within the community on how to develop such platforms. On the other hand, the cost of accessing space is decreasing, which lowers the barrier for new players to enter the market. Private companies have started to develop landing platforms and rovers to sell planetary access as a product. This leads to a need for reduction of mission costs and careful risk management to increase profits. In summary, we plan to investigate challenges and opportunities for the future of mobile planetary exploration. Next-generation mobility concepts were presented, which allow for investigating so far unreached planetary environments. A special focus was laid on the presentation of scientific targets and commercial opportunities.

The regular inspection of sewer systems is essential to assess the level of degradation and to plan maintenance work. Currently, human inspectors must walk through sewers and use their sense of touch to inspect the roughness of the floor and check for cracks. The sense of touch is used since the floor is often covered by (waste) water and biofilm, which renders visual inspection very challenging. In this paper, we demonstrate a robotic inspection system which evaluates concrete deterioration using tactile interaction. We deployed the quadruped robot ANYmal in the sewers of Zurich and commanded it using shared autonomy for several such missions. The inspection itself is realized via a well‐defined scratching motion using one of the limbs on the sewer floor. Inertial and force/torque sensors embedded within specially designed feet captured the resulting vibrations. A pretrained support vector machine (SVM) is evaluated to assess the state of the concrete. The results of the classification are then displayed in a three‐dimensional map recorded by the robot for easy visualization and assessment. To train the SVM we recorded 625 samples with ground truth labels provided by professional sewer inspectors. We make this data set publicly available. We achieved deterioration level estimates within three classes of more than 92% accuracy. During the four deployment missions, we covered a total distance of 300 m and acquired 130 inspection samples., Journal of Field Robotics, 37 (S 8), ISSN:1556-4959, ISSN:1556-4967

Planetary exploration robots encounter challenging terrain during operation. Vision-based approaches have failed to reliably predict soil characteristics in the past, making it necessary to probe the terrain tactilely. We present a robust, haptic inspection approach for a variety of fine, granular media, which are representative of Martian soil. In our approach, the robot uses one limb to perform an impact trajectory, while supporting the main body with the remaining three legs. The resulting vibration, which is recorded by sensors placed in the foot, is decomposed using the discrete wavelet transform and assigned a soil class by a Support Vector Machine. We tested two foot designs and validated the robustness of this approach through the extensive use of an open-source dataset, which we recorded on a specially designed single-foot testbed. A remarkable overall classification accuracy of more than 98% could be achieved despite various introduced disturbances. The contributions of the different sensors to the classification performance are evaluated. Finally, we test the generalization performance on unknown soils and show that the interaction behavior can be anticipated., IEEE Robotics and Automation Letters, 4 (2), ISSN:2377-3766

We present a tactile inspection approach for legged robots which allows for evaluating concrete deterioration in medium to large-sized sewers. The legged robot, to this end, executes a well-defined scratching motion with one limb on the sewer floor. Inertial and Force/Torque sensors implemented in specially designed feet capture the resulting vibration. The data is used with concrete deterioration estimates from professional sewer inspectors to train a support vector machine. We achieved deterioration level estimates within three classes of over 92% accuracy. The dataset, which we recorded during a field test campaign in the sewers of Zurich is made publicly available.

We present the outcome of a study on the energetic expenditure of quadrupedal gaits in the gravitational scenarios of Earth, Mars and the Moon. The study was performed in simulation on a fully controlled 30kg-class robot. We compared the mechanical power required for locomotion by using a static walking gait, dynamic gaits without full flight phases (trot, dynamic lateral walk) and dynamic gaits with full flight phases (running trot, pronk) at velocities up to 1 m/s. Additionally, we conducted a field test which compared the energetic expenditure and ground contact forces of a trot and running trot on a sandy terrain against the laboratory environment and the simulation results.Generally, gaits with full flight phases become increasingly efficient in reduced gravity scenarios. The study revealed that a running trot outperforms the gaits without full flight phases at forward velocities of 0.55 m/s on Mars and 0.4 m/s on the Moon. Executing a trot on the real robot showed that the energetic expenditure is 1.2-1.4 times higher on a coarse, heterogeneous sand compared to the lab environment. The field test revealed that the point feet design is not optimal for gaits with full flight phases on compressible soil due to high contact forces and increased ground penetration, which leads to stuck situations.

Simultaneous Localization and Mapping (SLAM) in unstructured planetary environments is a challenging task for mobile robots due to the appearance and structure of the environment. In urban and man-made scenarios, individual objects (e.g. cars, trees or buildings) are easily discernible and the visual appearance is likely to provide unique cues for the purpose of localization. Contrarily, planetary scenarios are often characterized by repetitive structures and ambiguous terrain features. To provide robust place recognition abilities in the context of submap-based stereo visual SLAM, we propose to utilize the gradient of elevation maps generated by Gaussian Processes (GPs). Visual features computed on GP Gradient Maps (GPGMaps) provide means for efficient place recognition, through encoding in Bag-of-Words vectors, and for SE(2) alignment to establish loop closure constraints in a pose graph. We evaluate the proposed SLAM system on relevant Moon-like environments through real data captured on Mt. Etna, Sicily.

The paper proposes a mission-aware energy planning strategy that highlights the weight of computations on the overall energy budget of possible future autonomous space robots for planetary exploration programs. The planning strategy addresses the problem of the limited power of such robots by extending some technical breakthroughs in the energy planning for autonomous terrestrial robots and attempting to apply them to the planetary exploration context. It saves energy by changing the trajectory and computation specific parameters during the mission as the robot batteries drain, accounting for a possible mission extension. It focuses on aspects of a space mission, such as its periodicity and uncertainty.

The main purpose of the project is to integrate sensors in the feet of the ANYmal, a quadrupedal robot capable of performing autonomous inspections in harsh environments, with the goal of performing sewer inspections included in the subTerranean Haptic INvestiGator project (THING). The objectives of the project have been divided between: the development of a haptic perception application which allows the ANYmal to differentiate between surfaces while scratching its feet; and the integration of sensors in the ANYmal’s feet to perform a water quality analysis in the sewer. To achieve the desired haptic perception, a foot mockup has been built to test the sensors selected for this application (piezo vibration sensors and an inertial measurement unit). The testing has provided promising results to validate the use of these sensors. After the results obtained with the foot mockup, these sensors have been integrated in the foot sole of the ANYmal and another a test has been performed, making the foot sole slide manually. The results of this test, again promising, have been verified using the Classification Learner App from the Machine Learning Toolbox of Matlab. Finally, a waterproof case has been designed to be attached at the frontal right foot of the ANYmal with the purpose of housing all these sensors. Since the sewer inspection also contains the water quality analysis, another case, which will be attached in the frontal left foot of the ANYmal, has been designed to house the selected sensors capable of measuring properties of the water such as the temperature and the electrical conductivity. The project has been concluded by performing a scratching motion test with the prototype attached in the feet of the ANYmal. Therefore, this project contributes to the THING project by giving the ANYmal the possibility of performing autonomous sewer inspections. Outgoing