2015年csmac演示风险知情干扰评估

J. Pierre de VriesCo-director, Spectrum Policy InitiativeSilicon Flatirons CenterUniversity of Colorado, Boulder May 12, 2015 Summary:Quantitative risk assessment can complement worst case interferenceanalysis and lead to more intensive coexistence of radio systems. Should a spectrum manager allow a new radio service if it might introduceinterference to an existing service? This question is at the heart of spectrumregulation. It has traditionally been answered by doing a worst case analysis that canlead to overly-conservative decisions. There is an alternative, though: risk-informedinterference assessment. Quantitative risk assessment has been used in other regulated industries for decadesbut has not yet been applied to spectrum management. A working group of the FCC’sTechnological Advisory Council (TAC) examined the potential of risk-informedinterference assessment last year, and recently published a paper on the topic (seereferences; I was a member of that working group). This column note the state of playand sketches the way ahead. Worst case analysis vs. quantitative risk assessment A worst case analysis considers the single scenario with the most severe consequence,regardless of its likelihood. However, there are many kinds of radio interference, andtheir impacts vary. A common but moderate effect may be more problematic overallthan the worst case. Selecting a single value of a single interference mode—typically aworst case—doesn’t represent reality accurately and can lead to false confidence thatthe resulting rules has averted any harm. For example, the rulemaking in the early2000’s to protect GPS from ancillary terrestrial transmissions in an adjacent bandfocused on out-of-band emissions, and the risk of adjacent band interference wasn’trealized until a decade later when LightSquared’s deployment loomed. A worst case approach is intrinsically conservative; it leads easily to rules that limitthe benefits of new services while giving incumbents more protection than they need. This approach made sense when spectrum rights were not in such great demand. It isnot tenable when high value services have to be packed in more tightly. There are exceptions where a conservative approach remains appropriate, such asservices where interruption is absolutely unacceptable and spectrum protection is theonly way to guarantee it. (Every incumbent will argue that it belongs in thiscategory.) In engineering practice, risk is often defined as the combination of likelihood andconsequence for multiple hazards, known as the risk triplet: What can go wrong? Howlikely is it? What are the consequences? By contrast, a worst case analysis focuses ona single scenario with very severe consequences, regardless of its likelihood. In general, the purpose of risk assessment is to provide evidence-based informationand analysis that can inform decisions on how to deal with risks and choose betweenoptions. In spectrum management, the risk is that of harmful interference and theselection is between various possible service rules (including the option of notallowing a new service at all). Applying this technique to spectrum yields risk-informed interference assessment. Quantitative risk assessment in other regulated industries Quantitative risk assessment has been used around the world for decades inregulated industries from finance to food safety, including cases where safety of life isparamount: •In 1967, Reg Farmer of the U.K. Atomic Energy Authority advanced the idea ofprobabilistic risk assessment for nuclear reactor safety analyses. The U.S.Nuclear Regulatory Commission adopted quantitative risk assessment in theSeventies. Its 1995 policy statement on probabilistic risk assessment (PRA)encouraged greater use of this technique to improve safety decision-makingand regulatory efficiency, and in 2009 it published guidance on the use of PRAto support licensee requests for changes to plant licenses. •The U.S. Environmental Protection Agency (EPA) uses risk assessment tocharacterize the nature and magnitude of health risks from chemicalcontaminants and other environmental stressors. The EPA first issued acancer risk assessment in 1976, followed by a series of guidelines based on a1983 risk assessment paradigm developed by the U.S. National Academy ofSciences. Risk assessment practices are now well established at the agencyand are widely used for public and environmental health protection. •The U.S. Food and Drug Administration (FDA) uses risk analysis to ensurethat regulatory decisions about foods are science-based and transparent. It hasdeveloped FDA-iRisk, a publicly accessible online tool to estimate the healthburden of microbial and chemical hazards in food. Risk assessment methods are also used by other U.S. government agencies anddepartments including the Offices of Management and Budget and Science andTechnology Policy; the Departments of Homeland Security, Health and HumanServices, and Transport; and the Federal Avia