SciBlogs

Archive April 2012

Pilots also optional? Robert Hickson Apr 29

No Comments

In my last post I noted the shift to increasing automation on the roads. We are also seeing this in the air. In the US legislative and budgetary initiatives are set to encourage greater use of unmanned aerial vehicles (aka Drones) in domestic airspace. Some fliers are in a tailspin about the dangers this poses to regular aircraft, because of the absence of collision avoidance systems in most planes and drones.

True, drones usually have a pilot controlling them at a distance, so they are not the same as self-driving cars. But it’s possible that pilot-less passenger aircraft aren’t far off (the belief that current jets can essentially fly themselves is a myth).

A range of organisations have already applied to fly drones in US airspace – the Air Force, Homeland, Security, Universities, and police departments. The latter make some particularly nervous because of the fear of invasion of privacy (rather than having weapon laden drones flying over a city). That’s a distraction from the main issue of the increasing capabilities of sensor systems generally and how they are used (it doesn’t really matter if they are in a small drone or a larger, manned aircraft). With drones I’d be more worried about the potential for accidents in a crowded environment.

However, two can play at the spy game, as Syrian protestors illustrated by keeping an eye on security forces with a camera attached to a model plane. I expect we’ll see more of this – there is already an active DIY drone culture.

New Zealand could potentially use drones for maritime surveillance. A Palmerston North company, Skycam UAV NZ, has developed drones for aerial surveillance.

Second hand military robots are also likely to be donated to police departments in the US — for surveillance and bomb disposal work (some police already buy them). With their widespread adoption by defence forces, I expect military robot manufacturers will look for new markets in police and domestic security applications.

DARPA is also sweetening the pot. In early April they announced a Grand Challenge to develop search and rescue robots. A key initiative they are looking for are systems of robots that can do a range of tasks and communicate with each other. If this Grand Challenge is as successful as their series of self-driving vehicle challenges, then we’ll see a lot more rapid innovation in the robot world.

I’ll end somewhat tangentially. This video is called ‘Robot readable world’, which is misleading because robots aren’t always involved. However, I found it eerily entrancing to see how computers are identifying objects and navigating streets.

Update 7 May: The American Civil Liberties Union have posted what they call “a nightmare scenario” of how police could slowly increase the use of surveillance drones to a point where they infringe privacy. The scenario relies on some future technological developments, such as face and gait recognition, improved visual analytics and coordination between devices, as well as broadening situations when surveillance can be used. The ACLU calls for regulations to be proactively employed to ensure such intrusive surveillance does not occur.  As I noted above, the privacy issue isn’t just about drones, but more generally about the power of sensor systems and the increasing number of ways they can be deployed (overtly or covertly).

Update 14 May: In Australia, the Victorian police are considering using drones for surveillance and searches. Chris Laidlaw’s Radio NZ Sunday Morning programme on 13 May  had a discussion about drones that is worth listening to.

Drivers optional Robert Hickson Apr 27

No Comments

Self-driving cars — such as Google’s (but also real car makers, like BMW, Audi and General Motors) — are still some years away from being available to the common or garden back seat driver. None-the-less Nevada has introduced regulations establishing procedures for their use on public roads, and California is also keen to do so. These seem more for economic rather than public safety reasons, since they hope to attract firms to set up shop and test there.

However, as both The Economist and Technology Review point out, more and more regular cars (not just top-of-the-line ones either) are getting some of the sensors and other features seen in autonomous ones. Self-drive creep, so to speak. It may take about 30 years for collision avoidance and other such systems to be the norm.

The US National Highway Traffic Safety Administration is encouraging development of automated systems (such as car to car communication) to enhance safety. A large European research programme, Safer Road Trains for the Environment, is also developing automated driving systems to improve safety and fuel efficiency.

Consideration is also being given in the US to require collision avoidance systems in all new cars in just a few years. While legislators there are already proposing installing black box recorders in vehicles from 2015.

As you’d expect, liabilty issues are a challenge to overcome, particularly in the US. Just look at what happened in response to unexpected acceleration in some Toyota models a few years ago. (There is still debate over whether there were electronic failures or human ones, such as having the wrong floor mats). Honda has also been sued for alleged failure in collision avoidance systems.

As vehicles continue to become more electronic and less mechanical a growing challenge will be to ensure that the different electronic systems communicate with each other (both within and between vehicles) as they become more complex.

However, concern over whether my Android car will be able to ‘talk’ to your iCar may be less of an issue than whether your Ford will even what to attempt to communicate with my Holden.

Greater automation also raises the issue of moral hazard. Drivers may take greater risks if they have electronic backup. Will we need a different driver licensing system?  Someone who has only learned to “drive” a semi- or fully autonomous vehicle is unlkely to have the skills to drive what we now consider to be a normal car. (This is going back to the future – if my memory serves me right, back in the day if you sat your driving test with an automatic transmission you weren’t certified to drive a manual car). Will car rental firms in the future charge a premium for non-autonomous vehicle?

One issue with greater driving automation that I haven’t yet seen raised is how autonomous driving will affect urban planning and the use of public transport. If you can read, watch TV, cruise the internet, etc while travelling to work, and driving becomes less stressful, the barriers to living further from work are likely to drop. Just as better roads and cheaper vehicles last century encouraged longer commutes and urban sprawl.

For some, their most productive work periods may be those travelling to and from the ‘office’. People may also be less likely to use public transport if sitting in your own autonomous vehicle is attractive. So do we want to see more urban sprawl as a cost of safer driving? Will “smarter cities” be undone by smart cars?

Update, 30 April: Technology Review have an item about MIT’s Media Lab designing a system to make sure autonomous vehicles can communicate with pedestrians to let them know it has seen them and they can safely cross the road.

Short term reforms Robert Hickson Apr 09

No Comments

Len Cook, President of the New Zealand Institute of Public Administration, recently criticised the New Zealand fetish for haphazard structural change in the public service. He noted that it undermines the need for continual learning and a culture of continuous improvement. It’s not that change may not be necessary, but that structural change should follow good analysis (as seems to be the case in Australia with their Moran Review).

He also noted the substantial societal, economic and environmental changes underway, and the need for the public service to have the capabilities, mechanisms and people to help address these challenges and opportunities. This picks up on the themes explored in Victoria University’s Institute of Policy Studies’ Future State report released in 2010. Future State noted the increasing instability and diversity in the world and New Zealand, and the consequent rise in unpredictability. They identified four main challenges facing the Public Service:

  • Affordability;
  • Complicated problems involving many players;
  • A more diverse and differentiated population; and
  • A world of faster, less-predictable change.

The recent restructures (and those proposed) don’t seem to take much if any account of these issues. Most of the changes are about reducing costs and improving the efficiency of the public service. As Len Cook noted, there is usually no evaluation after the fact about whether these changes have saved money and resulted in greater efficiency. Colin James commented that what is lacking when this Government talks about public sector reform  is effectiveness — are we getter better services and policy advice from these changes? Effectiveness matters more than efficiency to the public.

Some Ministers seem keen on remodelling the public sector along private sector lines and introducing greater top down command and control. However, this is often a selective exercise. A KPMG international study [PDF] noted that the public sector devotes considerably less emphasis on staff development than the private sector. As KPMG note, politicians favour the short term quick fix approach and ignore the longer term strategic changes that are required.

Types of Public Sector Reforms. Source: KPMG (2009) "Tough choices ahead"

Types of Public Sector Reforms. Source: KPMG (2009) "Tough choices ahead"

As the Future State papers note, as problems become more complex, top down approaches are less likely to succeed.

Reducing staff is the quick and easy option for reducing costs, but it doesn’t always make the organization more profitable or effective. Bloomsberg Businessweek has noted that many firms have realised that they have cut too deep into their workforce and are now rehiring. Zeynip Ton, writing in the Harvard Business Review, notes that retailers need to invest more in their employees. Unsurprisingly, there is not an inverse relationship between staff payroll and profitability. Some firms that pay their staff more and employ more people generate substantially more profit than lower cost competitors. Zeynip also notes that the quality of the work not just the number of jobs within firms is critical to success. James Surowiecki picks up these ideas in the March 26 edition of the New Yorker.

While the public sector isn’t directly comparable to the private sector, some of the basics are the same. Sure more routine work can now be outsourced or undertaken by computers. But the competence and quality of staff matters as much within government departments as in firms, as the recent criticisms about proposed changes at MFAT illustrate. Sadly, other departments don’t seem to be able to make the same point as effectively.

Reform in the public sector is necessary to meet emerging challenges and opportunities, but the government is taking a short term view that means we will not have the public service that we need.

Network-wide options by YD - Freelance Wordpress Developer