The Compressed Baryonic Matter experiment  at the future FAIR facility is dedicated to studies of QCD phase diagram at high baryonic densities and moderate temperatures produced in heavy-ion collisions. A very high collision rate up to 10 MHz is expected at CBM and continuous streaming readout is proposed.
It was shown recently by studies from the RHIC BES program that and the difference between of particles and antiparticles in the region of a few GeV are of great interest for understanding a pattern of the phase transition between quark-gluon and hadronic matter . Precision measurements of these observables in CBM experiment will be a significant step forward in exploration of the QCD phase diagram in the region of a = 2-5 GeV.
2. Simulation setup
10 AGeV Au-Au collisions are simulated with the UrQMD generator (version 3.4)  and the GEANT3 transport code. MDV, STS, RICH, TDR, TOF and PSD hits are simulated, then reconstructed into tracks and clusters and processed into reduced analysis tree within CbmRoot.
Three CBM detectors are of special importance for this performance study. The STS  and MVD detectors, located in the central rapidity region, are used for track reconstruction and identification as well as for event plane determination and centrality estimation. PSD detector  located in the forward rapidity region is used for spectator detection, event plane determination and centrality estimation. PSD geometry used in the simulations is: 44 modules 20 20 cm covering full azimuthal angle with three groups of modules. The beam pipe hole is 10 cm.
3. Strangeness reconstruction in CBM
The KF (Kalman Filter) Particle Finder algorithm  is used to identify V0 decays ( , , ) in the reconstruction. Invariant mass distributions for (Anti) and are shown at Fig. 1.
4. Centrality determination in CBM
Centrality is needed to obtain event classes for different impact parameter intervals. In CBM the centrality can be calculated with the PSD energy, the STS track multiplicity or a combined 2D distribution . For 1D distributions the fitting procedure with Negative Binomial Distribution is used . and parameters are obtained with Glauber Monte-Carlo model . For 2D distributions an iterative procedure is used for profiling, fitting, perpendicular profiling. It was shown  that by using a combined 2D estimator one can improve impact parameter resolution for central collisions (0-30% centrality). In the studies reported here we have used STS tracks multiplicity as the estimator for event centrality.
5. extraction procedure
Anisotropic transverse flow is the effect of azimuthal anisotropic particle production with respect to the reaction plane (1).
The scalar product (SP) method is used to extract the flow coefficients , eq. (2). In this method Q-vectors defined in (2) of subevents corresponding to 3 groups of PSD modules or STS subevents separated in rapidity are correlated with particle's unit vector. For the correction of the finite resolution a factor is used to obtain true values .
The invariant mass method to separate flow contribution of decaying particles from flow of combinatorial background is implemented: is calculated for each bin in invariant mass as well as the signal to background ratio. of the combinatorial background is estimated in the regions outside of the mass peak and of the signal is obtained with formula (3).
(3) where Bg and S are Background and Signal values, respectively, in the invariant mass distributions.
In experiment non-uniformity of detectors' acceptance leads to distortions of the Q-vector distributions. We utilize Q-vector Corrections framework  which implements corrections for these effects, such as gain equalization, recentering, alignment. In this study recentering correction is applied to the Q-vectors of each subevent.
A dedicated study was performed to check if eq.(4) is true or breaks due to non-flow effects and momentum conservation law. For simplicity, vectors defined in eq.(2) are scribed as taking into account that we used only 1st harmonic here.
Correlations between Q-vectors from PSD and STS subevents as well as correlations of Q-vectors and reaction plane angle are shown on Fig. 2. One can see that factorization is kept only for for mid-central collisions. The mixed harmonics procedure to calculate the resolution factor is necessary to obtain the correct value from observables.
Directed flow ( ) of and extracted for MC and reco particles is shown on Fig. 3. for two centrality classes (0-25% and 25-50%) are also shown on Fig. 3. Flow dependence obtained with SP method differs from model distribution both for and .
First performance studies with the UrQMD model and the CbmRoot detector response simulations were carried out. Collective flow methods are implemented to extract directed flow of baryons and mesons in Au+Au collisions with 10 AGeV energy expected at the future SIS100 accelerator in several centrality classes.
We observe the difference between model distributions and the ones obtained with PSD Q-vectors due to Q-vector factorization breaking.
This work is carried out with the financial support of FAIR-Russia Research Center. Also this work was partially supported by the Ministry of Science and Education of the Russian Federation, grant N 3.3380.2017/4.6, and by the National Research Nuclear University MEPhI in the framework of the Russian Academic Excellence Project (contract No. 02.a03.21.0005, 27.08.2013).