Try something like:
#Create some sample data
sample.data <- structure(list(Case = c(881L, 103L, 232L, 210L, 432L, 455L, 192L,
323L), MAC = c("10:B0:A0", "14:A1:FF", "13:A2:AB", "10:B2:C2",
"14:A1:FF", "13:A2:AB", "10:B2:C2", "13:A2:AB"), time_t = c(80L,
100L, 180L, 350L, 500L, 600L, 700L, 800L), Sensor = c(1L, 1L,
1L, 1L, 2L, 2L, 2L, 2L)), .Names = c("Case", "MAC", "time_t",
"Sensor"), class = "data.frame", row.names = c(NA, -8L))
#Split data into sensor 1 and sensor 2, then merge when MAC values coincide
summary_data <- merge(sample.data[sample.data$Sensor==1,],sample.data[sample.data$Sensor==2,],by="MAC",all=TRUE)
#Calculate elapsed time
summary_data$diff <- summary_data$time_t.y - summary_data$time_t.x
Note that if you have multiple repeated MAC values for both sensors, this will create a row for each combination of sensor 1 and sensor 2 - check that that's what you want.