Tham khảo tài liệu 'mobile robots perception & navigation part 17', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | Optimal Path Planning of Multiple Mobile Robots for Sample Collection on a Planetary Surface 631 except for Ơ8 A8 1 have been collected. One high-valued sample in particular 7 -4 9 is only collected in this eight-rover case. This makes sense since the sample is located close to the highest peak. The rover starting at x0 1 spends a long time climbing up to collect 27 and only collects four samples as a result. Note that the rover starting at x0 553 only collects three samples and has a very short mission time of sec although several high-valued samples are easily within reach. The algorithm does not try to distribute the task evenly which may be a hindrance to higher performance. However the results provide some insight on the nature of solutions to the SCP for multiple rovers. If we take the best values from the single rover V 43 two-rover V 84 four-rover V 136 and eight-rover cases V 169 the plot of mission value versus the number of rovers is sub-linear. If we take the worst values from the single rover V 37 two-rover V 72 four-rover V 136 and eight-rover cases V 169 the plot is again close to linear when the number of rovers is less than 4. Note that the spacing between the upper and lower bounds in the Mars terrain case is smaller than that of the flat terrain case. When examining the performance versus the number of rovers it makes sense that the graph is near-linear or sub-linear since once the best path for the first rover has been found there is not much room for improvement. In Fig 17 we observe that the flat terrain and Mars terrain results are near-linear or sub-linear and they are also very similar. The spacing between the upper and lower bounds for the Mars terrain is smaller than that of the flat terrain. This is due to the longer traveling times between samples in the Mars terrain case. The relatively close spacing of the rovers in our examples helps to ensure that nearly all the samples of interest are collected within the elapsed mission .