Merge 75ee629 into 65582b2

.gitignore

@@ -10,6 +10,12 @@ test/media/
 # media files generated by scripts
 scripts/media/
 
+# ipynb checkpoints in pythonscripts
+scripts/pythonscripts/.ipynb_checkpoints
+
+# idm acc trace files stored in lmfit
+scripts/lmfit/
+
 # Files generated by invoking Julia with --code-coverage
 *.jl.cov
 *.jl.*.cov

Project.toml

@@ -10,6 +10,8 @@ Combinatorics = "861a8166-3701-5b0c-9a16-15d98fcdc6aa"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 DelimitedFiles = "8bb1440f-4735-579b-a4ab-409b98df4dab"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
+ImageIO = "82e4d734-157c-48bb-816b-45c225c6df19"
+ImageMagick = "6218d12a-5da1-5696-b52f-db25d2ecc6d1"
 JLD = "4138dd39-2aa7-5051-a626-17a0bb65d9c8"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 PGFPlots = "3b7a836e-365b-5785-a47d-02c71176b4aa"
@@ -18,6 +20,7 @@ Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 Reel = "71555da5-176e-5e73-a222-aebc6c6e4f2f"
 Revise = "295af30f-e4ad-537b-8983-00126c2a3abe"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
+StatsPlots = "f3b207a7-027a-5e70-b257-86293d7955fd"
 
 [compat]
 Reel = "1.1.1"

scripts/experiments.jl

@@ -51,6 +51,24 @@ for i in 1:length(s)
         JLD.save("media/upper_$i.jld","m",m,"p",p,"md",md,"veh_id_list",veh_id_list,"ts",ts,"te",te)
 end
 
+#*****************Generate idm params for lmfit************
+s = scenarios_upper
+f = FilteringEnvironment()
+name = "upper"
+
+for i in 1:length(s)
+        veh_id_list = s[i][1]
+        ts = s[i][2][1]
+        te = s[i][2][2]
+        filename = "media/upper_$i.jld";
+        id_list,ts,te = JLD.load(filename,"veh_id_list","ts","te");
+        scene_list_true = replay_scenelist(f,id_list=id_list,ts=ts,te=te);
+
+        dirname = "$(name)_$(i)"
+        mkdir("lmfit/$(dirname)")
+        feat_dict = scenelist2idmfeatures(f,scene_list_true,id_list=id_list,
+                        scenario_name=name,scenario_number=i);
+end
 
 #************Docstring example code from multiscenarios_pf in helpers.jl*****
 # USEFUL to keep around for making tikz plots later on by rerunning this script
@@ -87,30 +105,93 @@ coll_mat_list = [coll_mat_idm,coll_mat_cidm,coll_mat_lmidm,coll_mat_pf];
 coll_barchart(coll_mat_list,filename = "media/coll_barchart_upper.svg");
 
 
-#********************Train upper test lower******************
-# We need to show a variability in the generated scenarios
-# So we need to combine the particle sets of different vehicles together
-# And then for the same set of vehicles in the lower merge i.e. test domain
-# We show significantly different driving behavior by sampling from the particle set
+#***********Histogram plot****************
+# Full example in docstring of Filtering/metrics.jl vel_distribution
+a = PGFPlots.Axis([veh_hist_true,v_hist_idm,v_hist_cidm,v_hist_lmidm,v_hist_pf,
+Plots.Command(raw"\legend{true,idm,cidm,lmidm,pf}")
+]);
+
+aidm = PGFPlots.Axis([veh_hist_true,v_hist_idm,Plots.Command(raw"\legend{true,idm}")],
+xlabel="velocity",ylabel="density",title="True vs IDM");
+
+acidm = PGFPlots.Axis([veh_hist_true,v_hist_cidm,Plots.Command(raw"\legend{true,c-idm}")],
+xlabel="velocity",ylabel="density",title="True vs C-IDM");
+
+almidm = PGFPlots.Axis([veh_hist_true,v_hist_lmidm,Plots.Command(raw"\legend{true,lm-idm}")],
+xlabel="velocity",ylabel="density",title="True vs LM-IDM");
+
+apf = PGFPlots.Axis([veh_hist_true,v_hist_pf,Plots.Command(raw"\legend{true,pf}")],
+xlabel="velocity",ylabel="density",title="True vs pf");
+
+PGFPlots.save("media/vhist_idm.svg",aidm)
+PGFPlots.save("media/vhist_cidm.svg",acidm)
+PGFPlots.save("media/vhist_lmidm.svg",almidm)
+PGFPlots.save("media/vhist_pf.svg",apf)
+
+#*******************Histogram2 plot for position trace****
+# Get the true position tracks for all upper scenarios and make histogram2
+cd("scripts");
+f = FilteringEnvironment();
+
+pos_x_master = Float64[];pos_y_master = Float64[];
+pos_x_master_idm = Float64[];pos_y_master_idm = Float64[];
+pos_x_master_cidm = Float64[];pos_y_master_cidm = Float64[];
+pos_x_master_lmidm = Float64[];pos_y_master_lmidm = Float64[];
+pos_x_master_pf = Float64[];pos_y_master_pf = Float64[];
+
+for i in 1:10
+    filename = "media/upper_$i.jld"
+
+    # Get position data for replay
+    id_list,ts,te = JLD.load(filename,"veh_id_list","ts","te")
+    scenelist = replay_scenelist(f,id_list=id_list,ts=ts,te=te)
+    pos_x_array,pos_y_array = pos_data(scenelist,id_list=id_list)
+    append!(pos_x_master,pos_x_array)
+    append!(pos_y_master,pos_y_array)
+
+    # Get position data for idm
+    scenelist_idm = scenelist_from_jld_idmbased(f,filename=filename,modelmaker=make_IDM_models)
+    pos_x_array_idm,pos_y_array_idm = pos_data(scenelist_idm,id_list=id_list)
+    append!(pos_x_master_idm,pos_x_array_idm)
+    append!(pos_y_master_idm,pos_y_array_idm)
+
+    # Get position data for cidm
+    scenelist_cidm = scenelist_from_jld_idmbased(f,filename=filename,modelmaker=make_cidm_models)
+    pos_x_array_cidm,pos_y_array_cidm = pos_data(scenelist_cidm,id_list=id_list)
+    append!(pos_x_master_cidm,pos_x_array_cidm)
+    append!(pos_y_master_cidm,pos_y_array_cidm)
+
+    # Get position data for lmidm
+    scenelist_lmidm = scenelist_from_jld_lmidm(f,scenario_name="upper",scenario_number=i)
+    pos_x_array_lmidm,pos_y_array_lmidm = pos_data(scenelist_lmidm,id_list=id_list)
+    append!(pos_x_master_lmidm,pos_x_array_lmidm)
+    append!(pos_y_master_lmidm,pos_y_array_lmidm)
+
+    # Get position data for pf
+    scenelist_pf = scenelist_from_jld_pf(f,filename=filename)
+    pos_x_array_pf,pos_y_array_pf = pos_data(scenelist_pf,id_list=id_list)
+    append!(pos_x_master_pf,pos_x_array_pf)
+    append!(pos_y_master_pf,pos_y_array_pf)
+
+end
+
+p = PGFPlots.Plots.Histogram2(pos_x_master,pos_y_master,zmode="log");
+PGFPlots.save("poshisttrue.svg",p) # Note that media/ is not part of the filename
 
+p_idm = PGFPlots.Plots.Histogram2(pos_x_master_idm,pos_y_master_idm,zmode="log");
+PGFPlots.save("poshistidm.svg",p_idm)
 
-#************LsqFit failed experimentation***************
-# model(t, p) = p[1] * exp.(-p[2] * t)
-# tdata = range(0,stop=10,length=20)
-# ydata = model(tdata, [1.0 2.0]) + 0.01*randn(length(tdata))
-# p0 = [0.5,0.5]
-# fit = curve_fit(model, tdata, ydata, p0)
+p_cidm = PGFPlots.Plots.Histogram2(pos_x_master_cidm,pos_y_master_cidm,zmode="log");
+PGFPlots.save("poshistcidm.svg",p_cidm)
 
-# function test_lmfit(d,p)   
-#         print("test_lmfit called\n")
-#         print("says:d=$d\n") 
-#         t = d["t"]
-#         return p[1]*exp.(-p[2]*t)
-# end
+p_lmidm = PGFPlots.Plots.Histogram2(pos_x_master_lmidm,pos_y_master_lmidm,zmode="log");
+PGFPlots.save("poshistlmidm.svg",p_lmidm)
 
-# test_tdata = []
-# for t in tdata
-#         push!(test_tdata,Dict("t"=>t))
-# end
+p_pf = PGFPlots.Plots.Histogram2(pos_x_master_pf,pos_y_master_pf,zmode="log");
+PGFPlots.save("poshistpf.svg",p_pf)
 
-# curve_fit(test_lmfit,test_tdata,ydata,[0.5,0.5])
+#********************Train upper test lower******************
+# We need to show a variability in the generated scenarios
+# So we need to combine the particle sets of different vehicles together
+# And then for the same set of vehicles in the lower merge i.e. test domain
+# We show significantly different driving behavior by sampling from the particle set