Hands-on robotic surgery is where the surgeon controls the tool's motion by applying forces and torques to the robot holding the tool, allowing the robot-environment interaction to be felt though the tool itself. To further improve results, shared control strategies are used to combine the strengths of the surgeon with those of the robot. One such strategy is active constraints, which prevent motion into regions deemed unsafe or unnecessary. While research in active constraints on rigid anatomy has been well-established, limited work on dynamic active constraints (DACs) for deformable soft tissue has been performed, particularly on strategies which handle multiple sensing modalities. In addition, attaching the tool to the robot imposes the end effector dynamics onto the surgeon, reducing dexterity and increasing fatigue. Current control policies on these systems only compensate for gravity, ignoring other dynamic effects. This thesis presents several research contributions to shared control in hands-on robotic surgery, which create a more natural motion for the surgeon and expand the usage of DACs to point clouds. A novel null-space based optimization technique has been developed which minimizes the end effector friction, mass, and inertia of redundant robots, creating a more natural motion, one which is closer to the feeling of the tool unattached to the robot. By operating in the null-space, the surgeon is left in full control of the procedure. A novel DACs approach has also been developed, which operates on point clouds. This allows its application to various sensing technologies, such as 3D cameras or CT scans and, therefore, various surgeries. Experimental validation in point-to-point motion trials and a virtual reality ultrasound scenario demonstrate a reduction in work when maneuvering the tool and improvements in accuracy and speed when performing virtual ultrasound scans. Overall, the results suggest that these techniques could increase the ease of use for the surgeon and improve patient safety.