120 likes | 264 Views
Speed Control and Scheduling of Data Mules in Sensor Networks. Presented by Justin Chester. Background. Sensor Networks Resource Constraints Multimedia Support Mobility Path Planning & Tour Planning Optimization & Heuristics Data Mules Unmanned Aerial Vehicle (UAV) Automobiles
E N D
Speed Control and Scheduling of Data Mules in Sensor Networks Presented by Justin Chester
Background • Sensor Networks • Resource Constraints • Multimedia Support • Mobility • Path Planning & Tour Planning • Optimization & Heuristics • Data Mules • Unmanned Aerial Vehicle (UAV) • Automobiles • Deer, Hikers, etc.
Example Applications • Structural Health Monitoring • Border Monitoring • Wildlife Tracking • Weather and Environment • Ferry for isolated networks
Data Mule Scheduling Problem • DMS is broken into three sub-problems • Path Selection • Speed Control • Job Scheduling • Focus on the 1-D DMS, speed profile and collection schedule that minimizes latency • Multiple Cases • Constant and Variable Velocity • Acceleration Constraint • Periodic Generation • Multiple Mules
Definitions • Job Scheduling • A job τi has an execution time ei and a set Ii of feasible intervals. A feasible interval I ∈ Ii is a time interval [r(I),d(I)], where r(I) is a release time and d(I) is a deadline. • Speed Control • A location job τi has an execution time ei and a set Ii of feasible location intervals. A feasible location interval I ∈ Ii is a location interval [r(I),d(I)], where r(I) is a release location and d(I) is a deadline location.
Definitions– cont. • Jobs must be executed within feasible intervals • Jobs can be simple or general • For an interval I = [r, d], |I| denotes the length d − r. We also define containment as follows: I ⊆ I′ if and only if r′ ≤ r and d ≤ d′ where I′ = [r′, d′].
Assumptions • A priori knowledge of: • Communication Range • Execution Time • Locations (Static) • Optional Speed and Acceleration Constraints • Ignore Curvature Constraint
Problem Definition • Instance: (L, J ), where • [0, L]: total travel interval on the location axis • J: set of location jobs, i-th job τi characterized as • Ii: set of feasible intervals • ei: execution time • Obtain time-speed profile v(t) • Map location to time to obtain induced job scheduling problem. • Determine v(t) such that job scheduling problem has a feasible schedule and total travel time is minimized
Basic Cases • Constant Speed – Maximum speed • Simple Jobs, O(n2) • General Jobs, linear programming • Variable Speed • Simple Jobs O(n3) • General Jobs, linear programming • Variable speed with Acceleration Constraint • NP-Hard • Propose 4 step Heuristic
Simplify: Convert general jobs to simple jobs • Maximize: Find maximum plateau speed and tight interval • Trim: Trim feasible location intervals of each location job. Calculate the time to be allocated to the remaining location interval. • Recursion: After previous step, all remaining jobs are split into two free intervals. Repeat step 2 on each.
References BARUAH, S. K., HOWELL, R. R., AND ROSIER, L. E.1993.Feasibility problems for recurring tasks on one processor. Theoret. Comput. Sci. 118, 1, 3–20. Sugihara, R. and Gupta, R. K. 2010. Speed control and scheduling of data mules in sensor net- works. ACM Trans. Sensor Netw. 7, 1, Article 4 (August 2010), 29 pages. DOI = 10.1145/1806895.1806899 http://doi.acm.org/10.1145/1806895.1806899 YAO, F., DEMERS, A., AND SHENKER, S. 1995. A scheduling model for reduced CPU energy. In Proceedings of the 36th Annual Symposium on Foundations of Computer Science (FOCS). IEEE Computer Society Press, Los Alamitos, CA, 374–382.