Military ground robots: a technological shift?

What are the chal­lenges of mil­it­ary land robot­ics today?

Dav­id Fil­li­at. Wheth­er mil­it­ary or civil­ian, robot­ics cov­ers dif­fer­ent tech­nic­al domains. There are sens­ory tech­no­lo­gies or mobil­ity tech­no­lo­gies, from a mech­an­ic­al or arti­fi­cial intel­li­gence point of view, etc. Impress­ive pro­gress has been made in recent years: Boston Dynam­ics’ robots, for example, are very suc­cess­ful in terms of phys­ic­al mobil­ity and are fully oper­a­tion­al. But a lot remains to be done if we con­sider the con­straints of the bat­tle­field or mil­it­ary oper­a­tions in a broad­er sense – both in terms of dis­cre­tion and robust­ness, we are only at the begin­ning of this story.

The greatest chal­lenges today relate to autonom­ous move­ment. Unknown and com­plex ter­rain has little in com­mon with a road where, broadly speak­ing, there are only ped­es­tri­ans and cars. In a mil­it­ary con­text the ter­rain is more rugged and var­ied. The need to be able to nav­ig­ate using GPS coordin­ates remains the same, but when it comes to the details of plan­ning the route the slight­est branch can be an obstacle that must be detec­ted, eval­u­ated, and man­aged – either by soft­ware or mech­an­ic­ally. So, there are addi­tion­al, high­er-level challenges.

Rather than humanoid (or dog-like) robots, we should ima­gine wheeled or tracked robots and the robot­isa­tion of equipment.

How­ever, design­ers ima­gine com­plete sys­tems, with each meth­od provid­ing solu­tions to cir­cum­vent cer­tain prob­lems. In robot­ics, for example, aer­i­al sys­tems can be sim­pler than ter­restri­al sys­tems. How­ever also brings a form of vul­ner­ab­il­ity, unless one can fly very high, far from oper­a­tions con­trol, which involves tech­nic­al choices. If we con­sider the on-board capa­city, ground vehicles (versus walk­ing or fly­ing) allow us to carry more energy and include more sensors. Each sys­tem has its own advant­ages and dis­ad­vant­ages; at the moment, there is no single ‘best’ solu­tion. We think more in terms of the dif­fer­ent applications.

Is full autonomy out of reach today?

On the bat­tle­field, cer­tainly. It is not in the com­ing years that we will see fully autonom­ous robots car­ry­ing out com­plex mis­sions. Robots will prob­ably appear along­side humans. Rather than humanoid (or dog-like) robots, we should ima­gine wheeled or tracked robots and the robot­isa­tion of equipment.

For example, we will see driv­ing aids in the form of con­voys, as is already being exper­i­mented with in the civil sec­tor with con­voys of semi-trail­ers, or in the form of self-driv­ing vehicles that take over from humans on open ter­rain. This lat­ter could allow the driver to leave con­trols and rest for an hour.

The main object­ive in all these tech­nic­al efforts is to save human lives. But this goes hand in hand with the search for oth­er, dif­fer­ent, more effi­cient ways of act­ing on the enemy. Robot­ics also comes into play in a con­text where new threats are appear­ing, which are them­selves dehu­man­ised, such as swarms of drones. This is there­fore a ques­tion of pro­tect­ing one­self, pro­tect­ing drivers and ground sol­diers, by mech­an­ising and auto­mat­ing cer­tain tasks.

Robot­ics is only a new stage in what is already a long his­tory. Dur­ing the Second World War, the Ger­mans designed the V1 to lim­it the loss of pilots, which are rare and highly qual­i­fied per­son­nel. Today’s armies reas­on in the same way with all troops, because in dis­tant con­flicts that are not fully legit­im­ate in the eyes of pub­lic opin­ion, a single human life lost has a great impact.

How­ever, we must not pre­tend that we are not wit­ness­ing changes. Fif­teen years ago, the applic­a­tions of robot­ics were still logist­ics, obser­va­tion, and sup­port. Today there is a switch towards to leth­al­ity. This effort is due in part to the per­cep­tion that the threats faced have evolved: in the short term, the per­ceived danger comes from swarms of small drones. But large coun­tries can devel­op lar­ger, stealth­i­er drones: the devel­op­ment of autonom­ous weapons sys­tems is a response to this threat, in an inev­it­able tech­no­lo­gic­al race.

What are the determ­in­ants of this tech­no­lo­gic­al race?

An essen­tial and new ele­ment, for me, is that we are talk­ing about dual tech­no­lo­gies, which are not being developed solely for the mil­it­ary. This is very dif­fer­ent from what happened in the nuc­le­ar or aero­naut­ic­al sec­tors, not to men­tion the many tech­no­lo­gies developed with­in the frame­work of the Amer­ic­an DARPA that were then dis­sem­in­ated to the civil­ian world (think of the Inter­net, GPS, etc.). Today, it can even be con­sidered that the civil­ian world (indus­tri­al­ists, major soft­ware play­ers) is a melt­ing pot for tech­no­lo­gies that could revolu­tion­ise this type of equip­ment. The mil­it­ary have under­stood this and are look­ing for syn­er­gies with the civil­ian world.

The mil­it­ary are look­ing for syn­er­gies with the civil­ian world.

This is the case when it comes to the abil­ity to ana­lyse sur­round­ings, where algorithms must be cap­able of pro­cessing huge volumes of data. This is where the civil­ian world may have an advant­age as the num­ber of sys­tems deployed that can acquire data is much great­er. Autonom­ous vehicles that car man­u­fac­tur­ers and equip­ment sup­pli­ers are cur­rently work­ing on must deal with decision-mak­ing, tra­ject­ory plan­ning and per­cep­tion prob­lems, which are all func­tion­al­it­ies that are also essen­tial for mil­it­ary use. Even if, as we have seen, the chal­lenges are great­er in a mil­it­ary con­text, and the data more dif­fi­cult to col­lect, the same tech­no­lo­gic­al build­ing blocks are being developed.