Despite the prevalence of this telerobotic surgery of humans, we still do not have the capability to service (refuel and repair) or assemble spacecraft on-orbit with telerobotic systems. Whether we intend to service active spacecraft on-orbit, or assemble new larger spacecraft on-orbit, the ability to do so with telerobotic systems could yield a new paradigm for spacecraft development and design. This would remove the constraints imposed on spacecraft designs by launch vehicles and enable the development of larger and previously-impossible space structures. Developing sufficiently capable and robust telerobotic systems for on-orbit operations is a technological challenge that and will require novel methods in order to yield a practical solution. My proposed research will begin with the augmentation of classical telerobotic systems with algorithms and approaches normally used in autonomous robotics. I will draw on my own research experience and familiarity in both of these subfields of robotics in order to develop new fundamental research in semi-autonomous telerobotics. My research plan includes: (1) incorporating more sophisticated sensing and motion-planning into telerobots, (2) developing algorithms and systems for multi-scale autonomy, and (3) using machine-learning to feed telerobotic experience back into the semi-autonomous systems. I believe we can develop a new class of robust semi-autonomous robotic platforms that will enable us to perform far more complex tasks than we could previously with telerobotic systems under high-latency telemetry.