On May 12, 2015, the Amtrak 188 skid down the Northeast Corridor, ultimately derailing after traveling more than double the recommended speed on the sharpest turns of the route. The seven-car train was filled 258 people total, and while it was not the most deadly crash that Amtrak has been associated with, it is a significant case study due to the initially ambiguous origin of the crash. Investigators shifted accusation from the mechanics of the train and railway to the engineer- Brandon Bostian. All were cleared of the most commonly attributable causes of crashes; the train was functioning properly to its programming, the rails were sturdy, and Bostian was not on his phone nor on any drugs. Thus, the investigation once again its focus on the human factors related to the incident- the way in which we as humans interface with the outside world and its objects, and the implications of it. Like the investigation, this paper will further analyze the human factors that played a role in the Amtrak 188 incident in order to identify what was a deficiency from a human factors perspective and to mitigate these in hopes to prevent future incidents like this one.
A great deal of the issues involved can be tied back to the situation and circumstances that did not support optimal functioning for Bostian, nor the effective operation of Amtrak 188’s alerting and tools. It was late at night, around 9:16 pm, when the crash occurred; it was dark, with the article mentioning that the headlights were the only source of light. Part of the Northeast Corridor’s notoriety was the curve’s lack of visibility, so these two factors coupled together contributed to a lack of environmental awareness, and thus situational awareness as well which affects the detection part of the decision-making loop. The other demands of the route greatly exacerbated the situation as well. The Northeast Corridor recently became one of the busiest and thus most challenging routes. It is mentioned that engineers riding these tracks are always “playing catch up (Article, Page 5)”, which places a great amount of stress on the engineers to make time. Given that Bostian’s “cab signals (Article, Page 7)” were unavailable during the first leg of his trip and he slowed to uphold safety, it comes as no surprise that witnesses reported that he appeared “frazzled (Article, Page 7)”. With the ingrained importance of time added to Bostian’s late arrival, it can be assumed that Bostian becomes more strained than his usual rides. In addition, these cab signals would have demanded a lower level of vigilance and would have mitigated the effects of missing a break. Since Bostian did not have these to rely on, the ride began to drain upon his attentional stores, as it operates cognitively as a limited capacity system. This would have inhibited his ability to detect and possibly interpret the situation at hand.
Due to the route’s ever-changing direction and orientation, engineers navigating this stretch of track are given no vigilance breaks. Often it is the case, even with cars, that you can decrease your vigilance acuity to some degree relatively safely when on a long stretch of straight road. However, given that the straightaways were quite short and infrequent, the engineers were not afforded this quasi-break. It did not help that the second leg of his couplet was just as draining. It is due to this demand for extreme accuracy that Karl Edler, a veteran engineer of the corridor states, “Even a moment’s inattention can be catastrophic… you’re constantly trying to think one step ahead” (Article, Page 5). It requires a large amount of attention, vigilance, and cognitive processing- all of which have their tolls magnified the longer that the engineer must sustain them. Pedestrians, railway workers, and drug addicts trespass with regularity, and it has become sport to hurl objects at passing trains- an activity referred to as “rockings (Article, Page 8)”. Because of these habits, train engineers must exercise acute vigilance throughout the entirety of the trip. As the article asserts, after a rocking the engineers have “no time to prepare (Article, Page 8).” As Bostian reported getting hit after a “hot track (Article, Page 10)” warning, he may have been able to anticipate it, but it most likely created an interruption from his mode of intense vigilance, and it was at the time that he began increasing speed. This interruption could have caused a lapse in his working memory, as Bostian quickly attended to the effects of the rocking, and perhaps forgot that he opened the throttle. This is where interruption recovery (Lecture 9, Slide 24) would have been vital, given that there was a sequence of steps to correctly maneuver that critical portion of the Corridor, and should Bostian have forgotten which step he was at, this lapse would contribute to his incorrect knowledge-based behavior (Lecture 9, Slide 2).
The dashboard of the Amtrak 188, an area that is designed to be quickly readable and assist in the detection stage of the decision loop, is noted to have its screens too close together- creating an issue with proximity clutter. Should Bostian have attempted to do a quick check of system statuses after the rocking, his eyes easily could have glazed over the train’s speed control, overwhelmed by all the information to process. It would have been difficult regardless of this to note the changes due to inattention blindness. Not only did Bostian need to check to see if any statuses changed, but also needed to recall in his memory what status for each criterion was within safe levels for where he was navigating. This lapse and inability to efficiently check statuses not only give issue to the detection and interpretation steps in the decision-making loop, but the action step as well, as Bostian was left with no feedback and no other form of error recovery after pulling the emergency brake.
In terms of personal attributes, Bostian himself was described as fairly young, at 32 years old. While the article does credit a great amount of passion and experience to Bostian’s early years, this does not qualify him to be classified as a seasoned and experienced engineer- especially given that the Amtrak 188 was a new model of train. It could have been that Bostian perceived himself as having the expertise and became overconfident (Lecture 8, Slide 39) and therefore overestimated his abilities. Based on the assertion the article made about how experienced engineers are able to successfully handle rocking occurrences and how Bostian could not, it can be asserted that Bostian is not yet experienced enough to handle these jarring and precarious situations, which require a reactionary response backed up with great experience as a skill-based behavior (Lecture 9, Slide 2). The accident was further worsened from what it could have been due to the lack of an effective error recovery method. Bostian attempted to pull the emergency brake, and under some circumstances, this may have worked. However, the Amtrak 188 had no other mechanism for when the issue has worsened beyond this point. The train thus passed the point of no return as it began traveling on own momentum, which not even a skilled locomotive engineer would have been able to save.
Another issue later on in the last part of the couplet: Bostian’s route had him continue through the Northern Philadelphia Station; while he did note ringing the bell as the train traveled through, it can only be wondered whether an added stop here would have provided a ‘mental reset’ and purposeful interruption for Bostian. During this stop, he would have the time to visualize, rehearse, and review the rest of the route and the T.S.R.B.s for the speed reductions necessary- as well as rest after a long stretch of acute vigilance. To relieve the effects of sustained attention (Lecture 3, Slide 2) and a vigilance decrement, breaks are not only important but necessary. It is unsustainable and grossly negligent for Amtrak to cut down the engineers’ break times between runs, and unfortunately, in this case, it completely prevented Bostian from getting a much-needed relief because he was behind schedule. Had he been given the previous average break, he still would have been able to take at least a thirty-minute break to relax and calm down from his “frazzled” state before needing to begin checking the Amtrak 188. The engineers’ union foresaw this issue before it even happened, citing concern over a decrement of focus and an increase in fatigue. Because the Northeast Corridor was one of the areas without the positive train control, or PTC system, Bostian, and other engineers needed to operate with insane accuracy and precision. Piloting at such a decisive level is not only emotionally taxing but cognitively as well.
Auditory clutter and interference very well could have been an issue as well. It is documented that the radio was on and chattering throughout the trip. While this uniquely would not have interfered with the visuospatial task of attending to the track ahead, this coupled with the “intermittent whine of the alterter (Article, Page 2)” could have easily caused Bostian to habituate to the sound of the alarms and thus decrease their salience. It is not mentioned whether there was an alert set up for an inappropriate increase in speed, but should there have been, its cue would have had less salience due to this habituation. Therefore, it is important for Amtrak in the future to consider better alarm management (Lecture 12, Slide 6) solutions.
The integration of a safety culture (Lecture 1, Slide 3) in the Amtrak organization was extremely lacking, occurring at the local level only and only observable through the use of safety goggles as far as the article mentioned. Amtrak expected their employees to pull off hours of painstaking precision, but denied them the much-needed breaks that would support their ability to do so. In the future, it would go far in the organization if Amtrak began restructuring the company to reward those that choose safety, and model this by supporting engineer’s ability to do so, like extending breaks, and thus implement the formation of a strong safety culture in their work (Lecture 8, Slide 37).
As previously mentioned, the PTC system was not implemented in the area where the Amtrak 188 crash occurred. Had it been installed as a failsafe, it could have caught this issue long before the situation grew to be out-of-control. The full implementation of PTC would not only save many other incidents like this from happening but also give engineers the personal benefit of being able to reduce their workload and vigilance to a safe degree. The PTC should only be used as a safety net, however, to avoid such happenings like the over-reliance seen in the Royal Majesty grounding incident. Technologies such as PTC should uphold trust, but not demand dependence (Lecture 12, Slide 12). Automation is not the answer to derailing accidents, as it does not perform well in novel situations and large complexity (Lecture 12, Slide 3); it is therefore still important to keep humans in this loop, as they can provide resilience, problem-solving, and reasoning. The aim should instead by at human-centered automation- to enhance human functioning rather than replace it. By doing so, we can take advantage of the U-shaped curve of automation: excellence in detection and acting.
If not able to implement the PTC program for whatever reason, another good measure to take would be to create a centralized control center for junctions, like how air travel has air traffic control centers. At particularly challenging junctions, such as the Northeast Corridor, speedometers will be set up and connected so that alerts can be sent for trains going beyond the maximum recommended speed, and thus the controllers will be able to radio the engineer to check up on the situation. Since this is at a delay and acts as an interruption, another safe option is creating an in-train display (Jack Cowan, Elden Arellano), like Waze for cars (below), where the engineer is able to quickly glance to check where they are, where they are going before they get there, and the recommended speed/speed limit for their current location. This should be paired with earcons (Lecture 6, Slide 17), to alert the engineer sparingly when they approach dangerous speeds. The use of an earcon rather than a visual alert will help keep interference at a minimum, as verbal and spatial information is encoding in parallel (Lecture 7, Slide 4). It will take less time to aggregate information, and thus help improve the speed-accuracy tradeoff (Jack Cowan; Lecture 9, Slide 7) while mitigating looking times and further distraction. Having it modeled after Waze’s interface (above) is beneficial because it uses pictorial realism (Lecture 4, Slide 4) and an egocentric view (Lecture 5, Slide 2). It also upholds the principle of the moving part (Lecture 4, Slide 11), and revolves around where the train is facing, heading, and currently at.
Amtrak has many directions they can go toward implementing new technologies, introducing safer procedures, and upholding more effective policies. By referring to human factor analysis, Amtrak and other corporations can focus their efforts more efficiently by attacking situations where technology goes right but all underlying issues and circumstances align to make it wrong. By changing factors and issues that leave errors open and vulnerable, they can design to close these gaps like in the Swiss Cheese Model (below). The fewer holes there are, the lower the probability they will all align in each train ride. They should focus on fatigue management and better risk assessment (Lecture 1, Slide 3), as well as all the above issues to diminish the probability of catastrophic events such as the Amtrak 188 accident from happening again.
