View Planning with Traveling Cost (Traveling VPP):

1
View Planning with Traveling Cost (Traveling
VPP) A Primal Dual Algorithm Based on the
Unified Problem Formulation ---
Pengpeng Wang, School of Engineering Science,
Simon Fraser University
I. Introduction Traveling VPP
3. Integer Program Formulation
Inspect object surface completely Realize
sensor actions by robot movements Minimize
total costs --- Sensing costs number of
views --- Traveling costs robot path length
Complete coverage requirement
Connection requirement
Unified formulation to facilitate the optimal
solution
2. Related Work 2-level Framework
4. Primal Dual Algorithm
First level view planning problem (VPP) ---
Minimize number of views Second Level
Traveling Salesman problem (TSP) --- Minimize
traveling cost
Modified first level to address sensing and
traveling (distance to the start position)
tradeoff
5. Performance Analysis
Counter example (two level framework performs
arbitrarily bad)
Views planned in optimal solution are close to
each other Two level framework --- First
level ignores traveling cost. Views planned can
be far apart --- Far-apart views result in
excessive traveling cost.
Same example (edge costs added)
Level I (modified sensing costs) --- v1, v2 and
v4 are chosen.
Level II (shortest path) --- Recover the optimal
solution