Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
222 changes: 222 additions & 0 deletions test/inertial/testIMUCases.jl
Original file line number Diff line number Diff line change
@@ -0,0 +1,222 @@

using Dates
using RoME
using CSV

df = CSV.read(@__DIR__() * "/testdata/IMUCase1.csv", DataFrame)
σ_a = 5e-4 # noise density m/s²/√Hz
σ_ω = 2e-6 # noise density rad/√Hz
Σy = diagm([ones(3)*σ_a^2; ones(3)*σ_ω^2])


a_b = SA[0.,0,0]
ω_b = SA[0.,0,0]

accels = map(SVector, zip(df.nax, df.nay, df.naz))
gyros = map(SVector, zip(df.nwx, df.nwy, df.nwz))
timestamps = df.time

pidx = range(1; step=100, stop=length(timestamps))
factors = map(zip(pidx[1:end-1], pidx[2:end])) do (fr,to)
r = range(fr,to-1)
RoME.IMUDeltaFactor(
SVector{3,Float64}.(accels[r]),
SVector{3,Float64}.(gyros[r]),
ones(length(r))*Ts,
Σy,
a_b,
ω_b
)
end

times = timestamps[pidx]

fg = initfg()
fg.solverParams.graphinit = false

foreach(enumerate(Nanosecond.(times * 10^9))) do (i,nanosecondtime)
addVariable!(fg, Symbol("x",i-1), Pose3; nanosecondtime)
addVariable!(fg, Symbol("v",i-1), Position3; nanosecondtime)
end

addFactor!(fg, [:x0], PriorPose3(MvNormal([1.0, 1, 1, 0, 0, 0], diagm(ones(6)*1e-3^2))))
# addFactor!(fg, [:v0], PriorPoint3(MvNormal([0.0, 0, 0], diagm(ones(3)*1e-3))))

# addFactor!(fg, [:x9], PriorPose3(MvNormal([1.0, 1, 1.0, 0, 0, 0], diagm(ones(6)*1e-3^2))))
# addFactor!(fg, [:v9], PriorPoint3(MvNormal([0.0, 0, 0], diagm(ones(3)*1e-3))))

for l in ls(fg, r"v\d")
addFactor!(fg, [l], PriorPoint3(MvNormal([0.0, 0, 0], diagm(ones(3)*1e-3))))
end

# Variable for bias
addVariable!(fg, :b, Position{6})

for (i,fac) in enumerate(factors)
frx = Symbol("x",i-1)
frv = Symbol("v",i-1)
tox = Symbol("x",i)
tov = Symbol("v",i)
if exists(fg, :b)
addFactor!(fg, [frx,frv, tox, tov, :b], fac)
else
addFactor!(fg, [frx,frv, tox, tov], fac)
end
end


debug = [:Iteration, :Change, " | ", :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :damping_term, "\n", :Stop]
debug = [:Iteration, :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :last_step_successful, "\n", :Stop]
stopping_criterion=StopAfterIteration(1000) | StopWhenGradientNormLess(1e-12) | StopWhenStepsizeLess(1e-12)

@time IIF.solveGraphParametric!(fg; stopping_criterion, debug, is_sparse=false, damping_term_min=1e-12, expect_zero_residual=false);

vnd = getVariableSolverData(fg, :b, :parametric)
vnd.val

## Test Case 2
df = CSV.read(@__DIR__() * "/testdata/IMUCase2.csv", DataFrame; limit = 3001)
# df = CSV.read(@__DIR__() * "/testdata/IMUCase2.csv", DataFrame)
σ_a = 0.037/60*sqrt(100) # noise density m/s²/√Hz - 0.037 m/s/√hr
σ_ω = 0.15*pi*60/180/3600*sqrt(100) # noise density rad/√Hz - 0.15 deg/√hr
Σy = diagm([ones(3)*σ_a^2; ones(3)*σ_ω^2])

a_b = SA[0.,0,0]
ω_b = SA[0.,0,0]

accels = map(SVector, zip(df.nax, df.nay, df.naz))
gyros = map(SVector, zip(df.nwx, df.nwy, df.nwz))
timestamps = df.time

pidx = range(1; step=100, stop=length(timestamps))
factors = map(zip(pidx[1:end-1], pidx[2:end])) do (fr,to)
r = range(fr,to-1)
RoME.IMUDeltaFactor(
SVector{3,Float64}.(accels[r]),
SVector{3,Float64}.(gyros[r]),
ones(length(r))*Ts,
Σy,
a_b,
ω_b
)
end

times = timestamps[pidx]

fg = initfg()
fg.solverParams.graphinit = false

foreach(enumerate(Nanosecond.(times * 10^9))) do (i,nanosecondtime)
addVariable!(fg, Symbol("x",i-1), Pose3; nanosecondtime)
addVariable!(fg, Symbol("v",i-1), Position3; nanosecondtime)
end

addFactor!(fg, [:x0], PriorPose3(MvNormal([1.0, 1, 1, 0, 0, 0], diagm(ones(6)*1e-3^2))))
addFactor!(fg, [sortDFG(ls(fg, r"^x"))[end]], PriorPose3(MvNormal([1.0, 1, 1.0, 0, 0, 0], diagm(ones(6)*1e-3^2))))

for l in ls(fg, r"v\d")
addFactor!(fg, [l], PriorPoint3(MvNormal([0.0, 0, 0], diagm(ones(3)*1e-3))))
end

# Variable for bias
addVariable!(fg, :b, Position{6})

for (i,fac) in enumerate(factors)
frx = Symbol("x",i-1)
frv = Symbol("v",i-1)
tox = Symbol("x",i)
tov = Symbol("v",i)
if exists(fg, :b)
addFactor!(fg, [frx,frv, tox, tov, :b], fac)
else
addFactor!(fg, [frx,frv, tox, tov], fac)
end
end


debug = [:Iteration, :Change, " | ", :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :damping_term, "\n", :Stop]
debug = [:Iteration, :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :last_step_successful, "\n", :Stop]
stopping_criterion=StopAfterIteration(1000) | StopWhenGradientNormLess(1e-12) | StopWhenStepsizeLess(1e-12)

@time IIF.solveGraphParametric!(fg; stopping_criterion, debug, is_sparse=false, damping_term_min=1e-12, expect_zero_residual=false);

vnd = getVariableSolverData(fg, :b, :parametric)
@test isapprox(vnd.val[1], [0.01, -0.02, 0.03, -0.01, 0.02, -0.03], atol=1e-3)


## Test Case 3

df = CSV.read(@__DIR__() * "/testdata/IMUCase3.csv", DataFrame)

σ_a = 0.037/60*sqrt(100) # noise density m/s²/√Hz - 0.037 m/s/√hr

σ_ω = 0.15*pi*60/180/3600*sqrt(100) # noise density rad/√Hz - 0.15 deg/√hr
Σy = diagm([ones(3)*σ_a^2; ones(3)*σ_ω^2])

a_b = SA[0.,0,0]
ω_b = SA[0.,0,0]

accels = map(SVector, zip(df.nax, df.nay, df.naz))
gyros = map(SVector, zip(df.nwx, df.nwy, df.nwz))

# accels = map(SVector, zip(df.ax, df.ay, df.az))
# gyros = map(SVector, zip(df.wx, df.wy, df.wz))

timestamps = df.time


pidx = range(1; step=100, stop=length(timestamps))
factors = map(zip(pidx[1:end-1], pidx[2:end])) do (fr,to)
r = range(fr,to-1)
RoME.IMUDeltaFactor(
SVector{3,Float64}.(accels[r]),
SVector{3,Float64}.(gyros[r]),
ones(length(r))*Ts,
Σy,
a_b,
ω_b
)
end

times = timestamps[pidx]

fg = initfg()
fg.solverParams.graphinit = false

foreach(enumerate(Nanosecond.(times * 10^9))) do (i,nanosecondtime)
addVariable!(fg, Symbol("x",i-1), Pose3; nanosecondtime)
addVariable!(fg, Symbol("v",i-1), Position3; nanosecondtime)
end

addFactor!(fg, [:x0], PriorPose3(MvNormal([1.0, 1, 1, 0, 0, 0], diagm(ones(6)*1e-3^2))))
addFactor!(fg, [sortDFG(ls(fg, r"^x"))[end]], PriorPose3(MvNormal([1.0, 1, 1.0, 0, 0, 0], diagm(ones(6)*1e-3^2))))

for l in ls(fg, r"v\d")[1:1]
addFactor!(fg, [l], PriorPoint3(MvNormal([0.0, 0, 0], diagm(ones(3)*1e-3))))
end

# Variable for bias
addVariable!(fg, :b, Position{6})

for (i,fac) in enumerate(factors)
frx = Symbol("x",i-1)
frv = Symbol("v",i-1)
tox = Symbol("x",i)
tov = Symbol("v",i)
if exists(fg, :b)
addFactor!(fg, [frx,frv, tox, tov, :b], fac)
else
addFactor!(fg, [frx,frv, tox, tov], fac)
end
end


debug = [:Iteration, :Change, " | ", :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :damping_term, "\n", :Stop]
debug = [:Iteration, :Cost, " | ", :Stepsize," | ", :GradientNorm," | ", :last_step_successful, "\n", :Stop]
stopping_criterion=StopAfterIteration(1000) | StopWhenGradientNormLess(1e-12) | StopWhenStepsizeLess(1e-12)

@time IIF.solveGraphParametric!(fg; stopping_criterion, debug, is_sparse=false, damping_term_min=1e-12, expect_zero_residual=false);

vnd = getVariableSolverData(fg, :b, :parametric)

@test isapprox(vnd.val[1], [0.01, -0.02, 0.03, -0.01, 0.02, -0.03], atol=1e-3)
Loading
Loading