### Table 8. Results (success ratios) for the tracking problem.

"... In PAGE 21: ... For the location problem, Table 7 shows the re- sults for 66 queries, and Figure 12 shows some examples. For the tracking problem, Table8 shows the re- sult of histogram backprojection and correlogram correction for the three test videos. These results clearly show that correlogram correction alleviates many of the problems associated with simple his- togram backprojection.... ..."

### Table 2: Physics-based tracking problem: Free and Fixed Parameters

2006

### Table 1: Example Race Track Problems. The results were obtained by executing Gauss-

1995

"... In PAGE 16: ...igure 1: Example Race Tracks. Panel A: Small race track. Panel B: Larger race track. See Table1 for details. the horizontal and vertical coordinates of the car apos;s location, and the second twointegers are its speeds in the horizontal and vertical directions.... In PAGE 42: ... We selected a state ordering for applying Gauss-Seidel DP without concern for any in uence it mighthaveonconvergence rate (although we found that with the selected ordering, Gauss-Seidel DP converged in approximately half the number of sweeps as did synchronous DP). Table1 summarizes the small and larger race track problems and the computational... In PAGE 43: ... After convergence of Gauss-Seidel DP,we compared these averages with the optimal expected path length obtained by the DP algorithm, noting the sweep after whichtheaverage path length was rst within 10 ;2 of the optimal. The resulting numbers of sweeps and backups are listed in Table1 in the columns labeled \Number of GSDP sweeps to optimal policy quot; and \NumberofGSDPbackups to optimal policy. quot; Although optimal policies emerged consider- ably earlier in these computations than did the optimal evaluation functions, it is important to note that this estimation process is not a part of conventional o -line value iteration algorithms and requires a considerable amount of additional computation.... ..."

Cited by 417

### Table 4.7: Resource utilization for the SIRF implementation of the bearings-only tracking problem on the Xilinx XC2VP125 device.

2004

Cited by 5

### Table 3.1 Example Race Track Problems. The results were obtained by executing Gauss-Seidel DP (GSDP).

1994

### TABLE III. GM_Plan GM_Learn Microcanonical Optimization Problem #Tracks Time #Tracks Time #Tracks Time States

1999

Cited by 5

### Table 1: Optimal Solutions to the Temperature Tracking Problem 5.1 Tracking Problem Let Td(~x) denote a desired temperature distribution inside the cavity. In our case, Td is the numerical solution to the dimensionless steady-state system (3.15)-(3.20) using FIDAP. We want to nd the right-wall temperature (or equivalently the Rayleigh number), so that the POD expansion given by

2001

Cited by 19

### Table 1. Correspondence between the target-tracking task and dynamics matching problem

"... In PAGE 4: ... The adjust- able dynamics a is the control of the information processing strategy. In this investigation the constrained dynamics c, the adjustable dynamics a, and the performance P are defined as in Table1 . The number of look-ahead steps k is selected as the adjustable dynamics a, which indicates how many steps ahead the motion should be predicted.... ..."

### Table 6 Results for race track problem with 21,371 states. Optimal solution visits only 2,248 states (Policy iteration is much slower than value iteration because exact policy evaluation has cubic complexity in the number of states, compared to the quadratic complexity of value iteration.)

"... In PAGE 21: ... Thus, only 2,248 states are visited by an optimal policy that starts at the beginning of the race track. Table6 summarizes the performance of LAO* on this problem. It shows the number of states evaluated by LAO* as a function of two different admissible heuristics.... In PAGE 23: ... LAO* expands the best partial solution graph each iteration until a complete and optimal solution is found. The last column of Table6 shows convergence times for this more efficient implementation of LAO*. (All timing results in this section are on a 300 MHz UltraSparc II.... ..."

### Table 6 Results for race track problem with 21,371 states. Optimal solution visits only 2,248 states (Policy iteration is much slower than value iteration because exact policy evaluation has cubic complexity in the number of states, compared to the quadratic complexity of value iteration.)

2001

"... In PAGE 21: ... Thus, only 2,248 states are visited by an optimal policy that starts at the beginning of the race track. Table6 summarizes the performance of LAO* on this problem. It shows the number of states evaluated by LAO* as a function of two different admissible heuristics.... In PAGE 23: ... LAO* expands the best partial solution graph each iteration until a complete and optimal solution is found. The last column of Table6 shows convergence times for this more efficient implementation of LAO*. (All timing results in this section are on a 300 MHz UltraSparc II.... ..."