### Table 1. User-space sensor observation types Continuous types

2005

"... In PAGE 5: ...nto process-level (PROC.*) and system-level (SYS.*). The observation type hierarchy shown in Table1 illustrates the type and granularity of information that is fed into TRAFEN. 3.... ..."

Cited by 7

### Table 1: Standard deviations of observations used for sensor calibration

### Table 1. The parameters of the Sensors.

1999

"... In PAGE 6: ... The NWS Forecaster predicts the availability of resources for each time-step of the measurements. Table1 shows the parameters of the Sensors. Next, we define a Bricks simulation under the second family of models mentioned earlier, employing the observed parameters of the real environment measured by Sensors, with cubic spline parameter interpolation, chosen because the interpolated value only depends on the local past (three time-steps).... ..."

Cited by 29

### Table 3: Class Sensor

1994

"... In PAGE 7: ... The attributes for env are given in Table 2. We consider the sensor (defined in Table3 ) to represent a typical hardware component, and its attributes largely reflect this. The exception to this is the Working attribute, which reflects the actual status of the sensor (as perceived by some hypothetical external omniscient observer), rather than providing some indication which is available to the using system.... ..."

Cited by 2

### Table 5 Validation Statistics by Sensor

2001

"... In PAGE 11: ... In 91% of the cases, the wake behavior was safely bound or not operationally significant. The same buffer categories are then broken out by sensor in Table5 . A general observation of interest is that the pulsed lidar had by far the highest percentage of valid wake files.... In PAGE 11: ... Figures 3 and 4 show the frequency of various magnitudes of negative buffer times for each sensor. As shown in Table5 and Figure 4, all hard exceedances were measured with the CW lidar. In these cases, the predicted demise was less than Taumin, and 75% of the observed demise times exceeded Taumin by less than 10sec.... ..."

### Table 1. Density of sensor nodes and desired radius

2004

"... In PAGE 7: ... Several factors affect the desired radius R, but we ex- pect that each sensor node can determine the appropriate radius R by observing its environment and estimating the density. Table1... ..."

Cited by 1

### Table 2: Preferable maneuver alternatives, maximum means of expected utility, and value of extra information related to different sensor information.

"... In PAGE 16: ... Next, the influence diagram is solved six times using different combinations of the sensors and numbers of observations. The summary of the results is shown in Table2 . For example, in the fourth case, the simulated decision maker receives ten observations from sensor 1.... ..."

### lable and observable and operates as a shift-register to delay the information coming from the preview- sensor to the disturbance input. Augmenting equations (1) and (3) together with (4) we obtain

### Table 2 shows other Monte Carlo simulations for the same satellite with different values for s, fit spans, and time intervals between sensor observations. Each Monte Carlo simulation in Table 2 contains 1000 trials. All the covariance derived sigmas are within the 95% confidence interval for the standard deviations and are thus good estimates of the standard deviations. The second row of Table 2 is the Monte Carlo simulation from Table 1. Since W = s -2 I, Eq. (7) implies that C is proportional to s2 and the covariance derived sigmas are proportional to s. The first three rows of Table 2 show this linear dependence on s. Rows 1, 4, and 5 of Table 2 show that the covariance

"... In PAGE 7: ... Real world sensor measurements may not be independent when the time interval between observations is short, in which case we would need to simulate the observations in a track as an auto-correlated time series instead of independent measurements. Rows 1, 6, and 7 of Table2 show that the covariance derived sigmas for the position errors are fairly constant for different fit spans, provided the number of sensor measurements remains constant. (There is a slight increase in sDv and a slight decrease in sDw as the fit span increases.... ..."

### Tables 4 and 5 summarize our main results. Note that the presented values for lifetime and depth are averaged across 20 di erent experiments for each network size. Further, the min and max columns for rm, rg and ri indicate the corresponding minimum and maximum performance ratios observed from those experiments. We make the following key observations for the smaller sensor networks:

in Efficient Algorithms for Maximum Lifetime Data Gathering and Aggregation in Wireless Sensor Networks

2003

Cited by 29