Struck-by accidents are one of the most deadly hazards found on
construction jobsites. But there is very few automated system to
calculate the hazardous zones.
Our target is to apply motion planning techniques to generate a safe
roadmap and produce a set of directions to the worker to avoid
collisions.
Ex-Ante Assessment of Struck-by Safety Hazards in Construction Projects: A Motion-Planning Approach
1. 1 / 26
Ex-Ante Assessment of Struck-by Safety Hazards
in Construction Projects: A Motion-Planning
Approach
Md Mahbubur Rahman, Triana Carmenate, Leonardo Bobadilla, Ali Mostafavi
Florida International University
October 20, 2015
2. Problem Definition
Problem Definition
Problem Statement
Solution
Experiment
2 / 26
Struck-by accidents are one of the most deadly hazards found on
construction jobsites. But there is very few automated system to
calculate the hazardous zones.
Our target is to apply motion planning techniques to generate a safe
roadmap and produce a set of directions to the worker to avoid
collisions.
3. Problem Statement
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
3 / 26
4. Motivation
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
4 / 26
Approximately 75% of struck-by fatalities involve moving bodies such
as trucks or cranes. (OSHA)
Two important factors affecting the level of safety hazards:
1) Sequence of activities and jobsite layout,
2) Movement patterns of workers and equipment
Complex, dynamic, and continuously changing nature of construction
jobsites like robotic path planning problem.
In our best knowledge no integrated evaluation of the construction site
layout is present in literature.
Possibility of using motion planning algorithms to avoid collision.
5. Problem Formulation in Motion Planning
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
5 / 26
Define the workspace, W, and the physical state space,X.
A moving body Bi is represented by a circle centered at, (Bi
x, Bi
y),
and its radius, r. B(t) denotes the position of the moving bodies at
time, t.
Obstacle region, Xobs, in the physical state space:
Xobs = {(q, t) ∈ X|A(q) ∩ (B(t) ∪ O) = ∅}. (1)
means some regions are shared between worker position and
obstacles; where q=configuration of a worker A and t is a particular
time parameter.
6. Problem Formulation in Motion Planning(Contd.)
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
6 / 26
The free space for the workers is denoted by Xfree = X Xobs.
The trajectory is piecewise linear functions of time, which means that
the translation applied to moving bodies can be written as
(Bi
x + c1t, Bi
y + c2t) for some constants, c1, c2 ∈ R.
xI ∈ Xfree is the initial state of the worker and XG ⊂ Xfree is the
final destination state.
7. Problem 1: Static Obstacles
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
7 / 26
8. Problem 1: Static Obstacles(Contd.)
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
8 / 26
Problem 1: Finding Safe Trajectories for Workers
Given an initial configuration, xI, a goal region ,XG, the set of static
obstacles, O, find an obstacle free trajectory, τ, such that τ(0) = xI
τ(1) ∈ Xgoal that is as far as possible from the static obstacles.
9. Problem 2: Moving Obstacles
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
9 / 26
10. Problem 2: Moving Obstacles(Contd.)
Problem Definition
Problem Statement
Motivation
Problem Formulation in
Motion Planning
Problem Formulation in
Motion
Planning(Contd.)
Problem 1: Static
Obstacles
Problem 1: Static
Obstacles(Contd.)
Problem 2: Moving
Obstacles
Problem 2: Moving
Obstacles(Contd.)
Solution
Experiment
10 / 26
Problem 2: Safety Evaluation of a Trajectory
Given the set of static obstacles, O, the motions of the moving
obstacles in time, B(t), and a trajectory, τ, evaluate the trajectory’s
safety policy and provide direction to the worker.
11. Solution
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
11 / 26
12. Assumptions
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
12 / 26
A geometric layout of the construction jobsite.
The trajectory of the moving equipment’s.
We cannot control the speed of the worker. But we can advice him to
stop at some point along his path to avoid collision.
13. Static Obstacle Avoidance
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
13 / 26
Create a safe roadmap/pathway for the workers to follow.
Maximum clearance roadmap as a form of Generalized Voronoi
Diagram which tries to remain as far as possible from static
obstacles(eg. excavation, building).
Each point on the path try to maintain equal distances from the edges
of Cobs.
14. Generalized Voronoi Diagram
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
14 / 26
Voronoi diagram decomposes the points into cells.
Generates a collection of Voronoi lines to separate the points on
space.
Prune the lines in intersection with obstacle edges.
Sample the wall to generate a clear and efficient roadmap.
15. Answer a Query
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
15 / 26
A worker wants to go from source point C to his destination point D.
Find out his roadmap.
The answer is to connect the nearest segments of voronoi diagram
with the source and destination.
Using Disjkstra’s algorithm to find out the shortest path.
16. Shortest Path is not the Safest
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
16 / 26
For static obstacles the shortest path is the safest. But we cannot
guarantee it until the consideration of moving obstacles.
So we must consider other paths following the voronoi diagram and
find out the number of collisions with moving obstacles on each path.
17. All Possible Path
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
17 / 26
All possible path can be obtained using DFS or BFS graph exploring
algorithm.
Here are some alternative paths found.
18. Solving Problem 2: Moving Obstacle
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
18 / 26
For moving obstacle the problem lies on the space time coordinate
system. (K. Kant and S. W. Zucker ’86).
Calculates the time when the obstacle is present on the workers’ path
and the amount of space on the trajectory occupied at that portion of
time.(Lavalle ’06)
19. Problem Represented in S-T Space
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
19 / 26
We must complete the total trajectory and reach the rightmost goal
line avoiding the moving obstacles.
Find out a way to avoid if an obstacle comes across.
20. Trapezoidal Decomposition
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
20 / 26
Decompose the free area in s-t system with a set of vertical line.
Connect the midpoints of the vertical lines and trapezoids to find out
the path to follow.
21. Adapt Trapezoidal Decomposition
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
21 / 26
An workers’ travelling path is a straight line with fixed slope in s-t
system(diagonal green lines).
A policy function π : (q, t) → {STOP, MOV E} based on s-t
system can be generated to direct the worker.
22. All Possible Paths
Problem Definition
Problem Statement
Solution
Assumptions
Static Obstacle
Avoidance
Generalized Voronoi
Diagram
Answer a Query
Shortest Path is not the
Safest
All Possible Path
Solving Problem 2:
Moving Obstacle
Problem Represented
in S-T Space
Trapezoidal
Decomposition
Adapt Trapezoidal
Decomposition
All Possible Paths
Experiment
22 / 26
As mentioned earlier the shortest path does not guarantee that it is the
safest. Rather the number of collision makes the worker slow down to
reach his destination.
We have to consider all other possible paths following the Voronoi line
segments to reach the destination and check hazards for each of the
trajectories.
24. Study Case I
Problem Definition
Problem Statement
Solution
Experiment
Study Case I
Study Case II
Conclusions
24 / 26
Following is a example s-t space graph for the shortest path. The
worker faces only one collision out of the two possible collisions.
Approximately 10 seconds delay for the worker as he lets the moving
body to pass.
25. Study Case II
Problem Definition
Problem Statement
Solution
Experiment
Study Case I
Study Case II
Conclusions
25 / 26
Followings are the two alternate paths those are not the shortest. But
we see the 1st path is the safest one as without collision. Also takes
very reasonable time for the worker to finish the path.
26. Conclusions
Problem Definition
Problem Statement
Solution
Experiment
Study Case I
Study Case II
Conclusions
26 / 26
We calculate safe trajectories that are as far as possible from static
obstacles while avoiding moving bodies (i.e., equipment).
Suggests some alternative paths with less hazards.
Next step is to process a heat map according to hazard rate on
different region of the construction zone.
A large scale planning is possible which can suggest alternate
construction activity sequences to minimize the risks.
Thanks!