Actuated Wingsuit Flight:
An Unfinished Journey

In 2009, we set out to make jet-assist­ed wing­suit flight sta­ble and con­trol­lable. Using first-prin­ci­ples mod­els and data from instru­ment­ed flights, we repli­cat­ed the known insta­bil­i­ty from ear­li­er attempts that used small tur­bo­jet engines attached to boots and explored a fly-by-wire approach. It is a sto­ry of ambi­tion and courage—and of a remark­able per­son whose pas­sion for flight brought joy, and then sorrow.

I first met Geof­frey “Geo” Rob­son in 2009, when we were, in a sense, inter­view­ing each oth­er. He was look­ing for a PhD posi­tion in sys­tems and con­trol, and I was look­ing for some­one to join the Fly­ing Machine Are­na team in my research group at ETH Zurich. Even though he had out­stand­ing cre­den­tials, I want­ed to see how well he could think on his feet. Sim­i­lar­ly, Geo want­ed to know if the cul­ture of our group was a good fit for his personality.

We sat down for what turned into an hour-long con­ver­sa­tion. By the end of it, I was con­vinced of two things: first, that he was tech­ni­cal­ly tal­ent­ed and would be a first-rate researcher; and sec­ond, that his per­son­al­i­ty would mesh with the col­lab­o­ra­tive, cre­ative atmos­phere we had built. There was a clar­i­ty of thought and a qui­et con­fi­dence about him that made me want to know more about what tru­ly drove him, so before we wrapped up, I asked, “If you could do research on any­thing you want­ed to, what would you do?”

A big smile spread across his face—one of those smiles that moves all the way to the eyes—and he laid out an idea that had clear­ly been with him for some time. An avid wing­suit fly­er, he want­ed to fig­ure out how to make wing­suit pilots fly lev­el using thrusters mount­ed on their feet.

This had been attempt­ed before—in 2005, Finnish wing­suit fly­er Visa Parvi­ainen had flown with small tur­bo­jet engines attached to boots. While he achieved short bursts of lev­el flight, the set­up was inher­ent­ly unsta­ble and could only be flown safe­ly at high alti­tude, where being far from the ground gave enough mar­gin to recov­er from the insta­bil­i­ty and deploy a para­chute. Geo’s idea was to go beyond this, build­ing a fly-by-wire con­trol sys­tem that would elim­i­nate the insta­bil­i­ty and make such flight tru­ly controllable.

His plan was con­crete and ambi­tious:
1. Col­lect wing­suit flight data—he would gath­er it him­self, instru­ment­ing his own flights.
2. Build a first-prin­ci­ples mod­el that explained the data.
3. Use the mod­el to design a con­trol sys­tem to sta­bi­lize flight.
4. Build it and fly it.

I pep­pered him with ques­tions. He answered each with clar­i­ty and care. By the time he fin­ished, I was con­vinced. “Let’s do that instead!” I told him.

Over the next year, we devel­oped and refined mod­els, test­ed assump­tions, and learned from each iter­a­tion. Our approach com­bined data that Geo col­lect­ed dur­ing his own wing­suit flights with first-prin­ci­ples aero­dy­nam­ic mod­el­ing, ground­ed in the same prin­ci­ples used to explain how air­planes and birds fly. To gath­er the data, Geo first had to design and build the instru­men­ta­tion him­self, cre­at­ing a sys­tem that could reli­ably cap­ture his posi­tion, speed, and ori­en­ta­tion through­out each flight.

The result was a paper titled “Lon­gi­tu­di­nal Sta­bil­i­ty Analy­sis of a Jet-Pow­ered Wing­suit.” In it, we ana­lyzed the flight dynam­ics of a wing­suit equipped with small jet engines, and our math­e­mat­i­cal mod­el was able to repli­cate the same insta­bil­i­ty that had lim­it­ed Visa’s ear­li­er flights. We then pro­posed con­trol strate­gies to over­come it, show­ing how human flight could be extend­ed beyond the lim­its of unpow­ered gliding.

Short­ly before head­ing home to South Africa for the hol­i­days, Geo came into my office, beam­ing, to tell me the paper had just been accept­ed for pre­sen­ta­tion at the AIAA Guid­ance, Nav­i­ga­tion, and Con­trol Con­fer­ence. It was one of his proud­est moments, and the joy in his voice made clear how much this work meant to him. He was also look­ing for­ward to see­ing his fam­i­ly and friends, and to wing­suit fly­ing in the moun­tains and val­leys of his beloved homeland.

On April 12, 2010, that future was cut short. Geo died dur­ing a wing­suit BASE jump from a cliff near Stel­len­bosch. He was just 31 years old. Los­ing him was devastating—for his fam­i­ly, his friends, and his col­leagues. For me, it was the loss of both a remark­able stu­dent and a close friend.

When the con­fer­ence came, I pre­sent­ed our paper in his place. It was the most dif­fi­cult talk I have ever giv­en; more than once I had to pause to steady myself. Stand­ing there, shar­ing the work he had poured him­self into, I felt the weight of his absence, but also the pas­sion that had dri­ven his pur­suit. Geo’s life was short, but he spent it pur­su­ing his dreams, and that is how I will always remem­ber him.