Haselton 



surge mode through large fixed control surfaces rather than with the control 

 system, he reduces not only the maximum control rates attainable, because of 

 the added drag, and the time to attain these rates, but introduces also motion 

 crosscouplings. 



The most important factor contributing to vehicle agility is the ability to 

 rapidly develop large control forces in the desired orientation. Not only does it 

 permit coping with fluctuating forces imposed by currents flowing about bottom 

 obstructions, but it greatly simplifies the control system designers' task and 

 that of the operator. From a control system consideration, it is desirable that 

 vehicle orientation changes be under the command of the human operator and 

 that vehicle stability be under the command of the control system. A good con- 

 trol system, for example, will make the operator oblivious to changing current 

 conditions and will depend on high-grade sensory inputs. This last requirement 

 places severe demands on the sensor designer, particularly those to sense sway, 

 heave, and surge components. Roll, pitch, and yaw data impose trivial demands 

 on the state-of-the-art. The operator's input device must be so orientated that 

 its motions are directly relatable in the most straightforward way to those of 

 the vehicle. That is to say, a pitch command, for instance, must be a pitch mo- 

 tion or force on the control input device. A good input device has zero reference 

 detents with sufficient breakout force to prevent undesirable control crosscou- 

 pling inputs by the operator, and damping. An example of such a device is 

 shown in Fig. 6. 



We have concentrated on adept vehicles. What if controlling the vehicle in 

 a precision manner is not the task but that accomplishing useful work is. This 

 implies that a prosthetic device of some description is to be operated from the 

 vehicle and will be used to extend man's desires to perform useful work func- 

 tions. There are two basic approaches to this problem; the first is to bottom or 

 otherwise anchor the vehicle physically with respect to the work area and then 

 proceed to concentrate on the operation of the prosthetic arm to accomplish the 

 task at hand. The second is to assume an untethered vehicle as the reference 

 platform. The first has the advantage of placing the operator in a situation 

 wherein he may leisurely go about his task, wherein the second places much 

 tighter requirements upon vehicle control and upon the prosthetic arm response 

 characteristics. There are, of course, situations wherein each can perform 

 where the other approach will not. For the vehicle at hand, it would appear that 

 the free flight approach would be most appropriate because the vehicle already 

 has imposed upon it requirements of control and stability nearly equivalent to 

 anchoring itself to the bottom. Once this decision has been made, it is possible 

 to design a control system which is prosthetic-arm-or-hand oriented. A first 

 requirement is that the dynamic response characteristics of the hand be signifi- 

 cantly faster than the vehicle. It has been demonstrated that the operator of an 

 arm is able to ignore vehicle stability in all six degrees of freedom and arm 

 response. In consonance with the principle of assigning all dynamic tasks to a 

 single operator, wherein possible, this arm control task presents an excellent 

 example. In the fulfilling of a specific task with the arm, such as picking up an 

 object resting on the ocean floor while performing a search mission, the oper- 

 ator should not have to make motion assignments between the vehicle and the 

 arm. His concern should be that of grasping the object with the hand. His or- 

 ders should be in terms of hand positioning control with no conscious concern 



310 



