Task Planning Clause Samples

Task Planning. The task planner has the objective to create a plan to achieve the current robot’s goal. The plan consists of a set of locations, chosen from a high level semantic map that represents different areas of the environment. The map produced by the mapping layer is annotated offline and then read by the Task Planner, which represents it as a graph structure. We developed a first version of the Task Planner for the early prototype of the system. using the dijkstra algorithm, in order to select the best path from the graph structure.. For the final version we decided to use the well known A* algorithm, instead, and integrated the Task Planner fully with the rest of the system. Figure 1 shows an example of a possible plan. ICT-FP7-600877-▇▇▇▇▇▇▇ Deliverable D5.6 In the Schengen barrier scenario, the robot will navigate in the Schiphol Airport, a very large area, which can pose problems for the motion planning modules. To handle this situation, we divide the environment map into a set of sub maps, which can be easily handled by the motion planners. As pre- viously said, the map is annotated offline and divided into a graph. In the annotation phase, the map is split into sub-maps, where each sub-map represents three linked nodes (n1, n2, n3) in the graph. This sets are not ordered, meaning that (n1, n2, n3) = (n3, n2, n1) . When navigating through a list of se- mantic nodes, the supervision system will choose a sub-map represented as (previous node in plan, current node in plan, next node in plan). Special care must be taken in two situations: when the robot is starting a plan (i.e. there is no previous node in the plan) and when the plan contains less than three nodes. In this situation we ▇▇▇- lyze the semantic map, and look for a node n prev where n prev edges(current node in plan) and n prev! = current node in plan. If n prev exists we select the sub-map represented by (n prev, current node in plan, next node in plan). If not, we look for a node n succ where n succ edges(next node in plan) and n succ! = current node in plan. The function edges(n) represents the list of nodes linked to node n in the semantic map. For example, the following sub-maps would be generated in the example from Fig. 1: During the example plan in the figure, the robot will choose the following maps: 1. Corridor B - Gate A - Corridor A 2. Gate A - Corridor A -Room A 3. Corridor A - Room A - Schengen Barrier The necessity of switching maps was already identified in year two. At that moment ...

Task Planning. One of the most effective methods and tools available to help supervisors protect the health and safety of their employees is a Pre-Task Analysis, or PTA. A PTA involves identifying the basic tasks of a job; determining any existing or potential hazards associated with each of the tasks; and then developing recommendations for eliminating or controlling each of those hazards. A PTA will be required each shift, before any major elements of work are performed, coordinating the efforts of subcontractors, trades people and other project personnel to ensure the common safety of all involved.