How to help swimmers win their races 

Mak­ing it to the podi­um can be a mat­ter of mere hun­dredths of a sec­ond. To max­imise the chances of win­ning medals, a great num­ber of fac­tors must be opti­mised. Rémi Carmigni­ani and his team’s research focus­es on so-called “row­ing” water sports: swim­ming, row­ing, canoe­ing, and kayak­ing. In these dis­ci­plines, ath­letes use an oar to move at the water/air inter­face – in swim­ming, this oar cor­re­sponds to the hand and forearm.

To vary their aver­age speed, swim­mers rely main­ly on the speed of exe­cu­tion of a peri­od­ic move­ment with their arms, called cadence. The high­er the cadence, the high­er the aver­age speed. To study this rela­tion­ship in swim­ming, a test is car­ried out in a pool: a pro­gres­sive 10 × 25 m with a start every three min­utes. The cadence and speed are mea­sured at each run. In swim­ming, speed evolves as does the cadence at high speeds¹.

The aim is to improve under­stand­ing² of the evo­lu­tion of speed as a func­tion of cadence and to relate it to the pow­er deliv­ered to the arms to move for­ward. For the time being, macro­scop­ic mea­sure­ments (aver­age speed, aver­age cadence, and total force in the so-called teth­ered stroke) have been made, but these will be fol­lowed by mea­sure­ments of the dynam­ic forces at the lev­el of the hands using strain gauges and of the kinet­ics of the move­ment with iner­tial mea­sure­ments. A gen­er­al mod­el to explain the dif­fer­ent regimes observed will then be pro­posed. This project is sup­port­ed by the EDF Foundation. 

Dur­ing the so-called active phas­es, the swimmer’s body is sub­ject­ed to addi­tion­al resis­tance forces that gen­er­al­ly increase the resis­tance in the water³. For exam­ple, dur­ing a wave, due to the defor­ma­tion of the body, the swim­mer must fight – in addi­tion to the pas­sive resis­tance of the unde­formed body – against an addi­tion­al force that is also qua­drat­ic with speed and depends on the shape of the wave. In swim­ming, in con­trast to oth­er “row­ing” sports, the swim­mer is both the shell and the oar.

A clas­si­fi­ca­tion and quan­tifi­ca­tion of the dif­fer­ent forms of active resis­tance for row­ing sports is underway.

This project start­ed in 2018 in the frame­work of the ANR NeP­TUNE (ANR-19-STHP-0004). The par­tic­u­lar­i­ties of swim­ming are the absence of equip­ment (no boat hull or oar, just the athlete’s body) and the pres­ence of the water/air inter­face mak­ing the use of equip­ment com­plex. This equip­ment can include iner­tial units (small water­proof box­es the size of a €2 coin con­tain­ing gyro­scopes and accelerom­e­ters that are attached direct­ly to the swim­mer) and force sen­sors. They can inter­fere with the swim­mer’s stroke and even affect the way they feel… 

One of the main sub­jects is the study of the swim­mer’s start and the opti­mi­sa­tion of his trajectory.

To sup­port these instru­ments, the INSEP pool has been equipped with twen­ty cam­eras (which film at a fre­quen­cy of 50 Hz). Ten are placed on the sur­face every five metres and ten under­wa­ter. The data cap­tured by these cam­eras is used by machine learn­ing pro­grammes to mon­i­tor the swim­mers. The force mea­sure­ments are used to explain the obser­va­tions by physics and to find para­me­ters to char­ac­terise the posi­tion of the swim­mer as well as his speed and acceleration.

One of the main top­ics is the study of the swimmer’s start, and the opti­mi­sa­tion of its tra­jec­to­ry. The start can be quite com­pli­cat­ed to opti­mise: the swim­mer starts out of the water on the pad, then breaks through the sur­face at a high speed rel­a­tive to the swim­ming speed, sinks under­wa­ter by about one metre before resur­fac­ing, and all this over a char­ac­ter­is­tic time of five sec­onds and a hor­i­zon­tal dis­tance of 15 metres. 

For the first part, try to under­stand the best take-off angle the swim­mer should have when leav­ing the pad. The aer­i­al tra­jec­to­ry that fol­lows until the swim­mer enters the water is less inter­est­ing from a physics point of view, as it can be described by the clas­si­cal equa­tions of free fall (air fric­tion is neg­li­gi­ble). What is impor­tant in this phase is the ori­en­ta­tion that the swim­mer will man­age to give to his body to enter the water by turn­ing around his cen­tre of mass. 

Remem­ber: by pro­pelling him­self, the swim­mer gen­er­ates addi­tion­al resis­tance! The speeds in this phase are high (near­ly 4 m/s). You should not trig­ger undu­la­tions too ear­ly at the risk of penal­is­ing your­self with greater resis­tance and final­ly brak­ing hard­er while try­ing to accel­er­ate. But be care­ful, wait­ing too long can also be prob­lem­at­ic as you risk sink­ing fur­ther and hav­ing to cov­er a greater dis­tance to return to the surface. 

In gen­er­al, in the front crawl, swim­mers trig­ger undu­la­tions when they reach the low­est point at the apex of their tra­jec­to­ry. This gives them zero ver­ti­cal veloc­i­ty. The sit­u­a­tion is dif­fer­ent in the back­stroke: they start in the descent phase, as they start with much low­er speeds.

The final objec­tive of this project, the the­sis work of Char­lie Pré­tot, who is a mem­ber of Rémi Carmigni­an­i’s team, is to help swim­mers opti­mise their tra­jec­to­ry over the first 15 metres (the length over which swim­mers are allowed to be under water in com­pe­ti­tion) at the start and at the turns. Should you go out a bit before the 15 metres? Maxime Grous­set, run­ner-up in the 100 m freestyle at the last World Cham­pi­onships, who has been mon­i­tored for about a year as part of this project, has shown that his best start occurs when he comes out of the water between 11 and 12 m. This project aims to help him improve his turn tra­jec­to­ry as well. Since the start of the project, more than 300 starts and turns have been analysed on about 20 swimmers. 

Since the begin­ning of the project, more than 300 starts and turns have been analysed on about 20 swimmers.

For the video obser­va­tions, the swim­mer is tracked using sev­er­al mark­ers on the skele­ton: head, hands, shoul­ders, ster­num, elbows, wrists, knees and ankles. A neur­al net­work was devel­oped to auto­mat­i­cal­ly track starts, turns and swim­ming for the first 15 metres.

In view of the open water swim­ming events that will take place in the Seine in 2024, sev­er­al projects are focus­ing on this dis­ci­pline. An analy­sis of the inter­ac­tion of swim­mers in an 80 m long canal in Cha­tou is under­way. The aim of this project is to under­stand using scale mod­els how, depend­ing on the posi­tion of the swim­mers, the wakes they cre­ate influ­ence oth­er swim­mers near­by. As in cycling, in this type of race, swim­mers can take advan­tage of the com­peti­tors’ suc­tion to save them­selves or even dis­turb anoth­er swim­mer. This work is being car­ried out by Bap­tiste Bolon, a research engi­neer in Rémi Carmigni­an­i’s team.

Anoth­er project also aims to pro­vide infor­ma­tion to the swim­mers of the French open water team on the expect­ed race con­di­tions: flow speed in the Seine and water tem­per­a­ture. For speed, this is the first time an Olympic race has been held in a riv­er since the rein­tro­duc­tion of open water races in Bei­jing in 2008. For the tem­per­a­ture, it influ­ences the equip­ment (neo­prene or fab­ric wet­suit) and on the peri­od of train­ing two weeks before the race. This work is being car­ried out with Mar­i­on Cocusse, a stu­dent at École des Ponts and is sup­port­ed by the EDF Foundation.

Isabelle Dumé