1
24-070
Perceptive Track Projection—Creating Context Sensitive
Path, Velocity, and Auxiliary Activity Projections for Use in
Autonomous Safety Intervention Systems
Robert Bissonette
CDC NIOSH, SMRD, Spokane, WA
Samir Sbai
CDC NIOSH, SMRD, Spokane, WA
ABSTRACT
Machine Situational Awareness (MSA) requires the abil-
ity to efficiently evaluate the probability of interaction of
objects within the environment, without creating false
alarms or ignoring materializing hazards. To do this, an
MSA system needs to assign context to observed object
activity and integrate as much information about object
identity, kinematics, behavior, and current state as pos-
sible. This paper describes how known object characteris-
tics, attributes, and activities (drilling, dumping, driving,
etc.) can refine interaction probabilities and help identify
rogue actors that introduce anomalous risks to the health
and safety of miners. Path projections need to incorporate
velocity and direction as a function of time while account-
ing for terrain, safety response, environmental factors, cur-
rent assigned activity, etc. This paper also describes the
methodology for properly prioritizing responses to place
the highest value on human safety.
MACHINE SITUATIONAL AWARENESS
(MSA)
The ISO 17757 [1] refers to situational awareness (SA)
as an overall concept of humans and highly automated /
autonomous (A/A) equipment working together. The stan-
dard covers many considerations for human factors and
roles in situational awareness. While there has been a great
deal of study and development on the human side of this
structure, the equipment contribution is in its infancy.
Given the absence of a formal definition, we found that the
existing terminology fell short in effectively expressing the
essence of our work. As a result, we introduced the term
“machine situational awareness,” which encapsulates the
equipment’s capacity to perceive its surroundings, assess
potential interactions within that context, and recognize
both existing and emerging hazards.
Humans have the innate ability to constantly evalu-
ate their environment and make millions of risk-reduction
decisions a day (some trivial, some lifesaving). The down-
side is that humans lose focus for a myriad of reasons such as
fatigue, boredom, distraction, etc. [2]. Machines equipped
with safety intervention controllers (parallel to machine
control) that employ MSA can potentially perform these
same functions without the shortcomings of humans, but
they will have to be nearly infallible to be trusted to oper-
ate in the presence of humans. Some events are tolerable in
a human-controlled environment that would not be in an
autonomous world (requiring mines to build new exclu-
sion zones, etc.). For a number of reasons, the expecta-
tion for equipment far exceeds that of humans [3]. Within
that paradigm, MSA must be substantially more robust
than humans. There are three critical elements to MSA, a)
redundant, consistent, and reliable perception of objects
24-070
Perceptive Track Projection—Creating Context Sensitive
Path, Velocity, and Auxiliary Activity Projections for Use in
Autonomous Safety Intervention Systems
Robert Bissonette
CDC NIOSH, SMRD, Spokane, WA
Samir Sbai
CDC NIOSH, SMRD, Spokane, WA
ABSTRACT
Machine Situational Awareness (MSA) requires the abil-
ity to efficiently evaluate the probability of interaction of
objects within the environment, without creating false
alarms or ignoring materializing hazards. To do this, an
MSA system needs to assign context to observed object
activity and integrate as much information about object
identity, kinematics, behavior, and current state as pos-
sible. This paper describes how known object characteris-
tics, attributes, and activities (drilling, dumping, driving,
etc.) can refine interaction probabilities and help identify
rogue actors that introduce anomalous risks to the health
and safety of miners. Path projections need to incorporate
velocity and direction as a function of time while account-
ing for terrain, safety response, environmental factors, cur-
rent assigned activity, etc. This paper also describes the
methodology for properly prioritizing responses to place
the highest value on human safety.
MACHINE SITUATIONAL AWARENESS
(MSA)
The ISO 17757 [1] refers to situational awareness (SA)
as an overall concept of humans and highly automated /
autonomous (A/A) equipment working together. The stan-
dard covers many considerations for human factors and
roles in situational awareness. While there has been a great
deal of study and development on the human side of this
structure, the equipment contribution is in its infancy.
Given the absence of a formal definition, we found that the
existing terminology fell short in effectively expressing the
essence of our work. As a result, we introduced the term
“machine situational awareness,” which encapsulates the
equipment’s capacity to perceive its surroundings, assess
potential interactions within that context, and recognize
both existing and emerging hazards.
Humans have the innate ability to constantly evalu-
ate their environment and make millions of risk-reduction
decisions a day (some trivial, some lifesaving). The down-
side is that humans lose focus for a myriad of reasons such as
fatigue, boredom, distraction, etc. [2]. Machines equipped
with safety intervention controllers (parallel to machine
control) that employ MSA can potentially perform these
same functions without the shortcomings of humans, but
they will have to be nearly infallible to be trusted to oper-
ate in the presence of humans. Some events are tolerable in
a human-controlled environment that would not be in an
autonomous world (requiring mines to build new exclu-
sion zones, etc.). For a number of reasons, the expecta-
tion for equipment far exceeds that of humans [3]. Within
that paradigm, MSA must be substantially more robust
than humans. There are three critical elements to MSA, a)
redundant, consistent, and reliable perception of objects