ABSTRACT

This paper presents a human factors perspective of the SVID (Simultaneous Vehicle and Infrastructure Design) process as advance by David Albright with the Alliance for Transportation Research and others. SVID embraces the system design approach that has evolved in the military and aerospace fields over the last half century or so. Human factors considerations involve interfaces with system elements and interactions with system operation. This paper summarizes those considerations and how they can be taken into account in the analysis, design and development process.

The SVID (Simultaneous Vehicle and Infrastructure Design) process involves consideration of the vehicle, infrastructure and the human operator in a classical systems analysis and design context. The approach of simultaneous consideration of the human, machine and environment is common and well developed in the military and aerospace fields and has commonly been referred to as human/machine systems studies which involve analysis, design and development. In military and aerospace applications this approach is carried out under a common authority such as the FAA, USAF, etc., and the considerations of human, machine and operating environment have been relatively tightly coupled. In public highway transportation, SVID must involve interaction between vehicle manufacturers and public road authorities, which traditionally has been rather loose and sporadic.

The consideration of Human Factors concerns both deterioration and enhancement of system performance. If system components, including the human operator, fail or operate inappropriately, then system performance degradation can lead to safety problems. On the other hand, enlightened system design, particularly in regards to human/machine interfaces, can enhance system performance beyond what might be considered merely adequate or acceptable. The objective of this paper is to present issues and considerations in system design and the design process when considering the interaction of the human operator, vehicle and infrastructure. This is a consciousness raising effort as opposed to a specific set of guidelines, but with some suggestions as to issues, resources and expertise that should be applied as part of the SVID design process.

Safety can be compromised in system design by a variety of causes including incorrect specification and implementation, component failures, and inappropriate human operator behavior. System components can be mechanized in hardware and software, and software complexity is of increasing safety concern in design and checkout. Some system components, such as vehicles, have traditional safety problems such as limit performance maneuvering, as well as potential ITS (Intelligent Transportation Systems) related safety problems associated with new innovative technology. Similarly, the human operator has traditional safety problems associated with impairment, training and inexplicable improper behavior (referred to as human error), as well as potential ITS related safety problems induced by poorly designed interfaces and task distraction and overload.

The SVID process must include consideration of driver, vehicle, and infrastructure characteristics and available technologies. This aspect of design is characterized by the conscious decision to identify control, guidance, navigation and hazard avoidance problems faced by the human operator and address them with specific system designs. It is important to insure that the impacts of all factors which can contribute to safety and performance are fully understood during system design. These factors include driver distraction, failure modes of both hardware and software, reliability of hardware and software, appropriateness of information/advice given to the driver and actions taken by the driver, ease of use, reduction in driving stress, and avoidance of driving situations with high collision potential.

Assessment of the impact of each factor involves establishing hypotheses about the impact, identification of measures of effectiveness and measures of performance that can be used to gather appropriate data, and determination of analysis methodologies to appropriately test the hypotheses. A safety impact analysis which provides this assessment can enhance the SVID process.

Human factors considerations are a key element of system design. The human/machine interface is a critical component in improving system safety and performance. The inclusion of sound human factors principles during the early stages of design can provide safety benefits from systems which are primarily designed for other purposes. Similarly, failure to apply human factors principles, or unintentional misapplication, can result in degrading the level of safety. In the following discussion this paper will attempt to define the role that human factors considerations can take in the SVID design process. We will explore possibilities for including human factors considerations in the SVID design process, in order to increase safety and enhance system performance.

Traditional system safety concerns involve system reliability and eliminating harm to the users, and include hardware and software issues in addition to human factors and human/machine interface considerations. Human factors considerations can go beyond safety, however, and can play a significant role in optimizing system performance and maximizing user acceptance. From a SVID view point, consider the human/vehicle/infrastructure system shown in Figure 1. This system structure depicts the human operator controlling a vehicle in the traditional highway environment. The driver perceives the highway environment including geometry, TCDs (traffic control devices) and traffic, and combined with perceptions of own vehicle motion, exerts control, guidance and navigation functions to achieve safety and performance goals. Given this structure, the SVID process should consider the human interfaces with the vehicle (e.g. controls, displays, seating, visibility) and the highway environment (e.g. visibility, sight distance, and TCD characteristics) in order to optimize the driver’s ability to maximize safety and system performance.

Now consider a more complex human/vehicle/infrastructure system as depicted in Figure 2 that imbeds the driver/vehicle component in a larger configuration such as a traffic management system. Although there is traditional emphasis on drivers and vehicles, advanced traffic management systems or commercial vehicle routing systems for example will have control rooms and professional operations personnel, and system hardware that requires maintainers and transit systems that involve travelers (e.g. bus passengers). Human/machine interfaces in the system will involve vehicle drivers as well as travelers (e.g. passengers), maintainers and system operators. The background, training and condition of personal transportation drivers and travelers can vary widely, while professional drivers, maintainers and system operators probably have been selected and trained for the skills they are expected to bring to the system. The Figure 2 example shows the broad transportation system human factors considerations that ultimately must be embraced in the SVID process. A common element are the human components and interfaces. With private passenger vehicles, we accept a wide range of operators with varying training and experience. With commercial vehicles we must consider the selection and training of professional operators and the requirements of travelers/passengers and cargo. Finally, traffic management and commercial vehicle routing systems bring into play control rooms and controls and displays that must be optimized from a human factors point of view in order to maximize system safety (i.e. minimize accidents) and performance (maximize flow and minimize congestion).

Although the emphasis in this paper is the human component of systems, it is recognized that the SVID process must consider the broad perspective of the interaction of humans, vehicles, systems and operations as suggested in Figure 3. Here it is implied that there is some union of human, machine, systems and operations that is relevant to a given problem, and the interaction of these considerations provides the context for analysis, design and development within the SVID process. The SVID process implies close coordination of a variety of disciplines (e.g. human factors, mechanical engineering, highway/traffic engineering, etc.) in an interdisciplinary context. It will require common communication and interaction between disciplines on projects and at technical meetings. Encouragement of the SVID process can be advanced through the interaction of disciplines and technical organizations such as the SAE, ASME, TRB, ITE, HF&ES, etc. Such multidisciplinary interaction will be important in advancing design objectives through the SVID process as discussed next.

GENERAL DESIGN OBJECTIVES

As mentioned above, overall design objectives from a human factors point of view involve maximizing both the performance and safety of transportation systems. In terms of performance, we would hope to lower congestion and pollution, improve ride and minimize deterioration of highway components, and improve the economy and efficiency of operations. From a human factors point of view, these system performance improvements should translate into system operator/user satisfaction due to improvements in travel time and mobility, reduction in hazards, and less frustration with system operations. These system goals might be as simple as optimizing ride and minimizing noise and vibration, improving TCD (Traffic Control Device) visibility and legibility and roadway line of sight, or optimizing road geometry to simplify vehicle guidance and control. In the realm of ITS, Advanced Traffic Management Systems (ATMS) and In-Vehicle Navigation Systems (IVNS) should simplify route finding and congestion avoidance, and generally reduce the frustration, wasted time and expense associated with navigation and route guidance. ACC (Automatic Cruise Control) and ACAS (Automatic Crash Avoidance Systems) should reduce driver workload and assist in hazard detection.

Additional human factors design objectives involve the learnability and usability of systems, and implied operator training requirements. Systems on personal transportation vehicles must work effectively with minimal operator training or experience, while training can be required for professional/commercial operators. System maintainability will also be critical in the efficient operation of new ITS innovations. When we consider human factors, selection, training and licensing considerations should be included in the SVID process. With personal vehicle drivers, we do not have the option of selection, so the design must account for a wide range of ability and experience, including particularly the novice and elderly driver categories. With commercial operators, we have control over selection and training which can accommodate more complex tasks.

The SVID process can impact system reliability and safety directly and indirectly. Certainly, system failures may contribute to vehicle accidents, especially if systems are not designed to be "fail-safe." All system failures are the result of hardware faults, software errors, data errors, operator errors, or design errors. The question, is: how do these kinds of failures manifest themselves in the human/vehicle/infrastructure and to what degree do they affect safety? For example, how reliable are onboard systems for augmenting vehicle stability or providing in-vehicle route guidance or information? What is their potential failure rate and how does failure influence the driving task?

In the engineering tradeoff between function, cost, and quality, the designer will often allocate certain responsibilities to the human operator. Professional operators with specified selection criteria and training requirements can be assumed to operate in a reasonably reliable manner. Unfortunately, the characteristics and abilities of the typical driver or traveler are largely unknown and uncontrolled. There are considerable opportunities for human error which can lead to safety concerns. A "safe" system must take into account the risk associated with various operator/user groups, and make provisions accordingly for operator/user mistakes and limitations.

Displays (format and content) – Vehicle/infrastructure systems will embody a variety of visual, auditory and possibly haptic (kinesthetic and proprioceptive sensation) displays designed to attract operator/user attention and expand situation awareness, provide warnings and control/guidance/navigation information, and command action under emergency conditions. The use of a display device determines the kind and amount of information which can be effectively and safely presented to the operator/user. Human factors considerations will determine acceptable displayed information format and content (e.g. brightness, color, loudness, frequency, etc.) that will have a major effect on the usability, safety, and overall appeal of a vehicle/infrastructure system. TCDs (Traffic Control Devices) present important external displays to the human operator, and there are plans to present such information with IVISs (In-Vehicle Information Systems). This information needs to be clear and unambiguous in its presentation. Visibility and legibility issues are important for external TCDs over a range range of weather and lighting conditions.

Controls (configuration and layout)–modern vehicles combine a range of on-board systems which can be complicated, and the human-machine interaction is often correspondingly complex. It is the job of the designer/engineer to manage this complexity with the careful application of human factors within existing physical and cost constraints. The designer is immediately faced with the problem of controls. How many controls are needed? Where are they located? Should physical controls be single or multi-function? How should the controls be labeled and what are appropriate motion stereotypes? Too many controls are intimidating, too few lead to confusion. Instrument panel real estate is often limited. Add-on devices must coexist with built-in vehicle instrumentation, heating/cooling vents, and air bags. Multi-function controls, can help organize feature-laden systems, but usability can suffer. Icons can be both a help and a hinderance.

Operator workload/distraction/attentional demands - Safety concerns demand special care in the design of the operator/user interface. This is particularly true when interaction with on-board systems (environmental control, entertainment, ITS) may distract from a more primary task such as vehicle control. The task is challenging because one must trade off the goal of providing sufficient amounts of useful information to the operator/user and requiring interaction, while at the same time minimizing the degree to which is distracted from other possibly more primary tasks. Divided attention ability and workload capacity will depend on the selection and training of the operator/user, and designs which achieve appropriate tradeoffs will limit the range of functions available to unsophisticated operator/users such as personal transportation drivers. Design tradeoffs will depend on the physical requirements of the operator/user interface, individual operator/user capabilities, and their overall familiarity with the system.

Biodynamics and occupant kinematics—Human response to vibration and impact are important considerations in vehicle ride and collisions. The human biodynamic response to vibration is a result of the interaction of vehicle tires, suspensions, and seating to inputs from the tire/pavement interface. Vibration can reduce comfort, impair vision, and cause injury due to both acute and chronic conditions. Collisions with other vehicles and roadside structures can cause serious injury, so that occupant packaging and restraint become a significant design issue.

System integration (control/display, warnings, etc) -.Vehicles currently have a number of on-board systems and ITS will only exacerbate this complexity. Control panels can be integrated. Form factor standards (e.g., DIN) can make a system easier to adapt to a variety of vehicle platforms. These endeavors are important because cost savings gained from careful integration of in-vehicle systems can help alleviate attempts at cost savings measures in other areas that have a negative effect on system reliability and safety.

Learnability, usability, maintainability - These issues can be handled to a large degree by selection and training for professional user/operators. For casual/unsophisticated operator/users (e.g. personal transportation drivers) acceptance of on-board systems will be largely determined by a person's initial experience using the system and its overall operability and performance. Unfortunately, the physical constraints imposed on the design of the human interface by its incorporation into the vehicle, make it very difficult to achieve a design which is both easy to learn and use. Applicable cognitive models which are familiar to broad segments of the population are virtually nonexistent. As a consequence, any particular design will be intuitive to some segments of the population and completely baffling to others. Of course the designer could lead the user through each operation with a generous dose of prompting, but the seasoned operator will find the interface tedious and inefficient. Training can help overcome these problems, but is not always possible and presents a definite obstacle to customer satisfaction.

Human operator population - Definitions of human operator populations to be accommodated run the gamut from selected and trained professionals to naive and unsophisticated drivers and travelers. Vehicle/infrastructure systems for personal transportation drivers and transit system travelers must be designed for the broadest range of users: young and old, novice and expert, men and women, all educational levels, and possible impairment (e.g. alcohol, fatigue). Some systems must be designed for the lowest common denominator, while others can assume operator/user selection and training. Given a targeted operator/user population, should the system contain only the most basic functions to avoid confusing the user? Should one seek the lowest common denominator of intellectual skill and experience among user groups to design in a shared cognitive understanding of the device? Should the IVHS designer provide the means to tailor the system for different individual preferences?

Safety and human factors design requirements must be substantially dealt with during the SVID design process. Often these considerations are relegated to design reviews, or even worse at the product testing stage, which is far too late in the design cycle to adequately account for safety and human factors concerns. Safety and human factors concerns and guidelines must be accounted for from the beginning conceptual design stage on through to prototype testing as alluded to above. Here, the design process is broken up into several stages, and safety and human factors issues are summarized for each stage.

This is the stage at which requirement specifications are developed. The operator/user population should be identified, and tasks and interfaces should be defined. The overall functional capability of the operator/user/system should be defined, and functions allocated between humans and machines. Given the target population and task definitions, selection and training requirements can then be specified. Typically for personal transportation drivers and transit system travellers, selection and training are difficult to control and systems must be designed for the lowest common denominator, and require minimal exposure for successful operation. Professional drivers, system operators (e.g. control room personnel) and maintainers can be selected and trained, although there can be significant costs associated with these activities that should be taken into account in the system requirements.

Questions will arise early on regarding human/machine interfaces (HFI), information processing, task cognitive and response time requirements, and the consequences of human error and system failure. Prototyping and simulation can help in resolving human factors questions while failure modes and effects analysis can provide a focus on safety consequences of human error and system failures.

Tradeoffs between hardware and software are decided at this stage, and details of the HMIs will begin to emerge. Error and failure tolerance should be analyzed, and fail/safe operational modes should be provided for.

A working prototype will allow for further safety and human factors analysis. Software verification and validation can be carried out at this point. Testing and evaluation should be carried out to establish consistency with the original specifications. The original specification can also be challenged at this point and updated if need be. Human factors testing should be carried out to evaluate HMI, training, useability and maintainability issues. Modifications should be made to correct HF deficiencies. Failure modes and effects testing can also be conducted at this point.

A final safety and HF review should be conducted at this stage. Durability and safety testing should be conducted with production prototypes. Final software verification and validation should be carried out on changes from the pre-production prototype.

Design evaluation procedures will be important in meeting safety and human factors requirements in the SVID process. Some analysis and simulation may be needed to establish and refine requirements and specifications. Prototype hardware and software should be evaluated for useability, training requirements and HMI characteristics at the earliest opportunity in the design cycle. Finally, final prototypes and production engineering models must be evaluated to verify and validate software, and to establish safety critical reliability. Some typical evaluation methods and approaches are as follows:

Analysis and Computer Modeling

Operator/user task requirements should be evaluated early in the design cycle. Task analysis and design procedures can be carried out during specification and preliminary design. The sensory, psychomotor and cognitive aspects of the operator/user task(s) should be reviewed as much as possible and the HMI requirements summarized. Discrete event simulation procedures can be used to evaluate process type Operator/System operations (e.g. navigation and traffic flow). Continuous time simulations and analysis procedures can be used to analyze continuous dynamic processes such as driver/vehicle guidance and control, biodynamic response, occupant kinematics, etc. Visibility and detection models can be used for analyzing the detection and recognition of TDCs.

Equipment prototypes should be set up as early as possible in the design cycle to permit testing for useability, training requirements, cognitive load, HMI issues, and workload and attentional demand. For cases where on-board systems will be operated in conjunction with other operator/user tasks, part task simulations should be carried out to make sure that all tasks can be completed safely under typical operational conditions. Workload and divided attention issues are particularly acute in the case of vehicle operators whose primary task is vehicle control and guidance. Subsidiary tasks must not significantly distract the driver, and part task driving simulation can be used effectively to investigate task demands.

A reasonable amount of evaluation can be carried out with real hardware/software in real world settings, although safety considerations will limit exposure to serious hazards. For example, instrumented vehicles can be run on test tracks and on public highways to examine operator reaction to system characteristics.

Data collection and monitoring of deployed systems should be considered. For example, the first application of a vehicle born system might occur in fleet trials which could include an autonomous data collection system. Data collection can also be built into large systems to monitor operations. These data collection systems will generate significant amounts of data, and some effort must be devoted automated data retrieval, reduction and analysis.

The SVID process is critical in dealing with the design of increasingly complicated vehicle/infrastructure systems. These systems include traditional concerns associated with the driver/vehicle/highway interface (tire/road interaction, roadside object encounters, TCDs, etc.), and the advent of innovative ITSs (information display, navigation, vehicle control, automated cruise control and crash avoidance). The system design process must embrace human factors principals in properly accounting for the human element in transportation systems.