After two years of commissioning beginning in 2009, much progress has been made in tuning the performance of LAMOST systems, including fiber positioning, dome seeing control, and optical alignments of spectrographs on the hardware side; and calibrations, data pipelines and data archiving on the software side. LAMOST is already producing useful spectra of bright objects, and can successfully observe targets as faint as r = 20 in the very best cases (and even fainter for emission line objects). LAMOST commissioning data have already produced a number of scientific results, including a search for metal poor stars, the discovery of new quasars and planetary nebulae, and mapping of the 2D stellar population pattern of the M31 disk.

In order to check the real instrumental performance and assess the feasibility of the science goals, the science committee of LAMOST came up with the pilot survey of LAMOST, which was conducted before the regular spectral survey. The pilot survey was launched on 2011 October 24, and ended in June 2012, which includes nine full Moon cycles and basically covers all good observing seasons during a year at the site. Each Moon cycle of the pilot survey was divided into three parts. Dark nights: five days sequentially before and after the new Moon; bright nights: five days sequentially before and after full Moon; grey nights: the remaining nights in the cycle. All time will be allocated to the survey, except the three nights attached to both ends of the dark nights, which will be used for technical purposes. In principle, dark nights will be used for faint objects with magnitude in the r (or V) band in the range [14.5 ~19.5], and bright nights should be for sources with magnitude in the r band of [11.5 ~ 16.5]. The exposure time for dark nights is 3 × 30 minutes, and for bright nights will be of order 3 × 10 minutes, depending on the distribution of the brightness of the sources. During the pilot survey, add-on projects were also available. Some special research projects, only needing a small number of fibers by reserving less than 10% of the fibers, can be proposed to the LAMOST operation center at any time, and will be scheduled once approved by the Science Committee of LAMOST. One of the proposals that meets the requirements and has been carried out is observing UV-excess targets, the science goal of which is to search for stellar mass black holes and neutron star binaries. 401 plates were observed and about one million spectra were taken during the pilot survey. In the early data release, more than 480 000 spectra with SNR ≥ 10 are kept. By using the combined pixel method to process the low quality data, 160 000 more spectra are gained. As a result, 640 000 spectra were released to the community after the pilot survey. The data volume of the LAMOST pilot survey proved that LAMOST is the telescope with the highest rate of spectral acquisition. The following is a comparison between LAMOST and other projects.The footprint of LAMOST is shown in Figure 1. 

                                    Fig. 1 Footprint of the LAMOST pilot survey

The LAMOST spectroscopic survey was launched in September 2012. Within 5 years, LAMOST will observe at least 2.5 million stars in a contiguous area in the Galactic halo, and more than 7.5 million stars in the low galactic latitude areas around the plane. The spectra collected for such a huge sample of stars will provide a legacy that will allow us to learn detailed information on stellar kinematics and chemical compositions well beyond that provided by SDSS/SEGUE. Such a dataset will allow for studies of the Galaxy before the Gaia era starts.