| <-
Layered Safety Approach
- Applying a far-field approximation, defining V_0=hE_x as the voltage gradient in the slot, and assuming k_0 h≪1 gives the equation for the direction of the radiated energy. Solving this equation for varying values of Θ and Φ generates a plot of the antenna coverage pattern. Figure 5 shows a sample antenna coverage pattern plot. Different antenna configurations produce different plots and multiple antennas may be used in concert. For example, multiple antennas could be arranged linearly, as two-dimensional arrays, meandering lines, etc. [13].
Historically, the traditional approach to robotic system safety has been to physically separate humans from robots with physical fences, light curtains and other keep-out zones. As robotics has evolved, however, we see a more collaborative environment where robotic systems and people are working in the same workspace. This presents a much more challenging situation with regards to safety than physically separate keep-out zones. Approaches to meeting this challenge will be multi-layered across different physical components and across different spans of time with sensing that is multi-modal across different sensing technologies. Ultimately integrating the input from this array of sensors becomes a multi-criteria decision making problem.
A hazard analysis is one of the first steps when designing safety systems. The mechanics of a hazard analysis for robotic systems in the abstract is beyond the scope of this paper, but robots have been in the workplace long enough for OSHA to give us guidance as to the hazards and types of accidents associated with these systems [7]. With robotic systems we are typically concerned with striking, ejecting, pinching, pushing, tripping, trapping, crushing and cutting. Some example accidents from the OSHA guidelines involving robotic systems include: a worker removed an imperfectly formed piece from a conveyor and the worker's back was forced against the robot; after adjusting a metal shaving machine, an operator was caught between the machine and the removal robot; a welding robot went functionally awry and its arm flung a worker against another machine; a worker removed the cover of an assembly robot to retrieve a fallen part and caught his hand in the robot's drive train; a robot's arm functioned erratically during a programming sequence and struck the operator; a fellow employee accidentally tripped the power switch while a maintenance worker was servicing an assembly robot; and an operator entered a robot's work envelope during operations and was pinned between the back end of the robot and a safety pole. In summary, robotic systems can be quite hazardous. The challenge is to have people and robotic systems working together safely in the same
workspace.
Safety standards around robotics have evolved to address collaborative robotics. ISO13482, for example, “specifies requirements and guidelines for the inherently safe design, protective measures, and information for use of personal care robots.” Personal care robots are most certainly examples of collaborative robotics.
The most recent standard for industrial robots, ANSI/RIA R15.06-2012, describes four modes of collaborative interaction between people and robots. The first is safety-rated monitored stop where the operator may interact with the robot when it is stopped, but still under power. Automatic operation may resume when the human leaves the collaborative workspace. The next mode is hand-guiding operation where the operator is in direct contact guiding the robot while it is moving. The next mode is speed and separation monitoring where the robot speed is reduced the closer an operator comes to the hazard area and a protective stop is issued when the operator is in potential contact. The fourth mode is power and force limiting to the extent that incidental contact between robot and person will not result in harm to person. Experience has shown this mode is more complex than it seems. Entering and then stopping in a person’s path when the person is not expecting it can be both a tripping and impact hazard. Again, there are lessons to be learned from existing deployments of collaborative robotics.
→
|
