We propose to overcome the limitations of wheeled surface rovers by combining recent advances in ball-shaped soft-robots based on tensegrity structures (a tension network of rods and cables), with a hopping mechanism based on cold-gas thrusters. The ball-shaped tensegrity robot with a payload suspended at its center can be collapsed into a small deployment volume, be light-weight, and navigate difficult terrain. In addition to be capable of rolling dynamically (by actuating its cables) and to survive significant landing impact shocks while protecting a delicate payload, we propose to dramatically increase the mobility of this design by adding a simple gas thruster located near the center payload, so that the robot can quickly cover a 1 km distance over a series of hops, while protecting its payload at the center of its structure. Once it has hopped close to its destination, the robot can roll to its exact target. For this concept, we propose to study three technology areas: 1) Mobility - allowing safe and accurate positioning of the robot with a combination of cold-gas thrusters and precision rolling, 2) Perception - placement of sensors to allow for robust navigation, and 3) Autonomy - control algorithms allowing for robust hopping and precision rolling mechanisms that can automatically overcome unexpected obstacles and failures. The innovative tensegrity-based surface probes have much potential in accurately deploying small payloads with speed, robustness and at costs unmatched by today's systems. Tensegrity probes can facilitate an intriguing low-cost exploration mission profile: 1) A tensegrity probe can be squeezed into a small cargo volume then automatically spring away from a base rover; 2) Once released, the tensegrity probe can use a simple gas thruster to make a series of hops towards a final destination. 3) An active tensegrity structure can be used to orient thruster. 4) Compliance of structure allows probe to ``bounce'' on impact, protecting the payload; 5) The probe can then reorient itself from landed position without addition reorientation hardware and efficiently move from landed position to perform sensor measurements and deliver payload; 6) It can survive significant falls and is resistant to being stuck, simplifying route planning and allowing for more aggressive maneuvering and increased autonomy. While tensegrity-based probes have the potential to dramatically increase performance and science return of robotic planetary probes, significant early stage technology development is still needed to enable and evaluate this technology for planetary missions. We propose to develop three key areas to enable this mission concept. These areas will be defined within a reference mission of safely delivering a 1 kg payload a distance of 1 km on the moon. 1. Define thruster and hopping profile to safely deliver payload: The unique tensegrity design offers a number of different ways that the mission goal can be completed. We propose to investigate what thruster hardware will be needed for such a mission to succeed. In addition we will investigate several forms of hopping, ranging from long hops, cushioned by reverse thrusters to many short hops. 2. Determine controls needed to orient thruster and navigate effectively: The design of this probe calls for significantly different controls than for a standard gimbaled thruster. Our thruster will be located within the probe and may be partially or fully oriented by changing the shape of the tensegrity. We propose to research how to develop controls that allow successful completion of the reference mission. 3. Characterize performance using simulation and hardware prototype: For a robot probe to have broad applicability, it will need to handle many challenging terrains, such as hills, and craters. We propose to simulate the tensegrity probe in a number of terrains and simulate its mechanical properties of hardware to show the practicality of this concept.