Term
| What is the Closed-Loop Online Method (Ultimate Sensitivity)? |
|
Definition
| 1. Turn off I and D (Tc->inf; Td=0). 2. Increase slowly until system is marginally stable. 3. The value of Kp when the system is marginally stable is Ku. |
|
|
Term
| What is the Modified Closed-Loop Online Method (1/4 decay)? |
|
Definition
| 1. Turn off I and D action. 2. Apple a step input. 3. Adjust Kp until the step response has the 1/4 decay property. 4. This value of Kp is called Km. |
|
|
Term
| What are the three approaches for CT to DT? |
|
Definition
| A. Numerical integration/differentiation. B. Pole-Zero matching. C. Hold equivalent. |
|
|
Term
| What is the solution for Numerical Integration/Differentiation? |
|
Definition
| y(kT)=y((k-1)T)+ int from {(k-1)T} to {kT} {[-ay(t)+au(t)]dt} |
|
|
Term
| What is the forward rectangular rule? |
|
Definition
| T[-ay((k-1)T)+au((k-1)T)] and in z: Gc(z)=Gc(s)|s=(z-1)/T |
|
|
Term
| What is the backward rectangular rule? |
|
Definition
| T[-ay(kT)+au(kT)] and in z: Gc(z)=Gc(s)|s=(1-z^-1)/T |
|
|
Term
| What is Tustin's method (Trapezoid Rule)? |
|
Definition
| (T/2)[-ay((k-q)T)+au((k-1)T)-ay(kT)+au(kT)] and in z: Gc(z)=Gc(s)|s=(2/T)[(z-1)/(z+1)] |
|
|
Term
| What is the solution for Pole-Zero Matching? |
|
Definition
| 1. Convert real poles s=a to z=e^(aT) 2. Convert complex poles s=a+jb to z=re^(jtheta) 3. take finite zeros and step 1 and 2 them. 4. infinite zeros are mapped to z=-1. 5. Gc(z)|z=1 |
|
|
Term
| What is the solution for hold equivalence? |
|
Definition
|
|
Term
| What is PID and what is the practical PID controller? |
|
Definition
| P=Kp.e(t) I=Kp.[1/(Ti)] int from 0 to 1 {e(t)dt} D=Kp.Td(de(t)/dt) or P=Kp I=Kp/(ti.s) D=Kp.Td.s For practical PID controller: Kp.Td.s.E(s) is replaced by E(s).(Kp.Td.s)/(1 + td.s/N) for the derivative term (D in PID) |
|
|
Term
|
Definition
| If there exist enough processor cycles to execute all the tasks. |
|
|
Term
| What is meant by a feasible scheduling algorithm? |
|
Definition
| The possibility to allocate the tasks so that all the deadlines are met. |
|
|
Term
| What is meant by an optimal algorithm? |
|
Definition
| One which will produce a feasible schedule if it exists. |
|
|
Term
| What is an cyclic executive (CE) approach to scheduling? |
|
Definition
| An execution schedule for tasks created before run-time and a program controls execution of tasks according to this schedule. |
|
|
Term
|
Definition
| Tasks are inserted at the end of a ready list and then execution taken from the beginning of the list but can not be preempted. |
|
|
Term
|
Definition
| Round Robin. Similar to FIFO, but tasks are preempted when its time runs out and then re-inserted at the end of the ready list. |
|
|
Term
|
Definition
| Rate monotonic. For a set of independent tasks, they are assigned fixed priorities based on its period. The shorter the period, the higher the priority will be assigned. |
|
|
Term
|
Definition
| Deadline monotonic. Like RM but the shortest deadline is considered instead of period. |
|
|
Term
|
Definition
| Earliest deadline first. This ensures a task with the earliest deadline to be scheduled first. The ready task with the earliest deadline has the highest priority at any point in time. |
|
|
Term
| What would be potential problems for dependent tasks? |
|
Definition
| 1. Deadlock. 2. Priority Inversion. 3. Transitive blocking. |
|
|
Term
|
Definition
| Original Priority Ceiling Protocol. Each task has a static default priority assigned and a dynamic priority that is inherited from blocking higher priority tasks. Each resource has a static priority ceiling value. A task can only lock a resouce if its dynamic priority is higher than the ceiling value of any currently locked resource. |
|
|
Term
|
Definition
| Immediate priority ceiling protocol. Like OPCP but the dynamic priority is raised to PC(R) as soon as there is a successful lock. |
|
|
